Page 1
Kim, D.-H., Lee, H. and Kwak, J. (2017) Standards as a driving force that
influences emerging technological trajectories in the converging world of
the Internet and things: An investigation of the M2M/IoT patent
network. Research Policy, 46(7), pp. 1234-1254.
(doi:10.1016/j.respol.2017.05.008)
This is the author’s final accepted version.
There may be differences between this version and the published version.
You are advised to consult the publisher’s version if you wish to cite from
it.
http://eprints.gla.ac.uk/154254/
Deposited on: 21 December 2017
Enlighten – Research publications by members of the University of Glasgow
http://eprints.gla.ac.uk
Page 2
1
Standards as a driving force that influences emerging technological trajectories in the
converging world of the Internet and things: An investigation of the M2M/IoT patent
network
Dong-hyu Kim, Heejin Lee, Jooyoung Kwak
Abstract
While standards are said to create windows of opportunity in facilitation of technological
convergence, it is not clear how they affect technological trajectories and strategic choices of
firms in the face of convergence and in the process of catch-up. There is little research on the
relationship between standards and technological trajectories, particularly in the age of
convergence. This paper investigates how standards shape the emerging M2M/IoT
technological trajectory and influence convergence in terms of technological importance and
diversity. We, firstly, found that standards are a driving force of technological convergence.
The second finding is that 3GPP standards assume a crucial role in setting the boundary
conditions of the M2M/IoT technological systems. Third, we identified strategic groups and
strategic patents that centered around the M2M/IoT trajectory. Forth, standards serve as an
important factor in the process of creating a new path for catch-up firms (e.g. Huawei).
These findings make contributions to innovation and standards studies by empirically
examining the relationship between technological trajectories and standards. Furthermore,
they clearly cast light on ongoing cooperation and competition along the M2M/IoT trajectory,
and offer practical implications for catch-up strategies.
Key words: M2M/IoT, standards, technological trajectory, catch-up
Page 3
2
1. Introduction
The Internet of Things (IoT) has emerged as a new technological paradigm that deconstructs
industrial silos, disrupts existing competitive relationships in the global markets, and opens
a window of opportunity for new entrants. It is broadly expected that billions of everyday
objects will be connected to the Internet within a decade and drastically affect people’s daily
life. In 1998, the term Internet of Things was coined by Kevin Ashton who described it as “a
standardized way for computers to understand the real world” (Schoenberger, 2002). It
embodies a vision where everyday objects with embedded sensors and actuators are
seamlessly connected to the Internet, enabling “ubiquitous computing”, a concept proposed
by Mark Weiser in 1988. Through the embodiment of everyday life, ubiquitous computing
helps people to find new ways to mobilize socio-technical assemblages (Galloway, 2004).
A variety of IoT services, based on an immense amount of data extracted from
numerous sensors, are possible only when the seamless connectivity of things is ensured.
Lack of communications among heterogeneous IoT networks and incompatibility among
multi-layered system components are considered a technological bottleneck, constraining
the current IoT evolution. A wide range of standardization activities is required to solve this
problem. Such standardization efforts, accompanied with changes in architectural
knowledge regarding relational properties of existing information technology (IT) systems,
can have crucial impacts on a competitive landscape in the converging world of the Internet
and things. Previous literature points out that there are a positive association between
standards and technological importance (Rysman and Simcoe, 2008) and a negative
association between standards and technological diversity (Blind, 2004). Technological
importance and diversity are essential concepts to understand technological convergence
(Cho & Kim, 2014). Yoffie (1997a) explained that standards played a critical role in digital
convergence. However, it has not yet been clarified how standards affect technological
convergence, particularly with respect to the relationship between technological importance
and technological diversity. This vacancy in the literature will be addressed in this paper.
The selection of certain standards among competing technologies creates lock-in
effects, and thereby influences subsequent innovations in a path-dependent manner (Arthur,
1994; Shapiro & Varian, 1999a). This path-dependence of technological evolution highlights
the relationship between standards and technological trajectories for future innovations.
Regarding the relationship between technological trajectories and formal standards, prior
studies focus on theoretical explanations between these two concepts (Metcalfe & Miles,
1994) and the identification of technological trajectories using patents (Verspagen, 2007). Yet
to the extent of our knowledge, there is no previous research that empirically shows how
formal standards affect emerging technological trajectories. This research gap also needs to
be filled.
Among different evolutionary trajectories of IoT (e.g. RFID, WSN, and M2M1), M2M
is often regarded as the main pattern of IoT evolution at the present stage (Chen, Wan, & Li,
2012). There are currently a variety of standards-setting organizations and consortia (e.g.
ITU, IEEE, IETF, ETSI, 3GPP, oneM2M, OIC, IIC, Thread, and AllSeen Alliance), which have
1 According to 3GPP (3rd Generation Partnership Project), M2M (machine-to-machine, also known as machine-
type communication (MTC)) is defined as data communication among devices without the need for human
interaction.
Page 4
3
made considerable efforts to accelerate the development of M2M/IoT standards
(Schneiderman, 2015).2 During the period of technological paradigmatic transition, a
strategic decision as to which standards to choose critically affects firms’ future performance
due to issues of compatibility. Bound to a common fate in the era of an emerging
technological paradigm, firms with similar technological capabilities are likely to form
strategic alliances and interact in a cooperative and competitive manner (Gnyawali & Park,
2011; Nohria & Garcia-Pont, 1991). Catch-up firms, in particular, are sensitive to the setting
of standards in the time of a techno-economic paradigmatic shift, as new standards bring
technological challenges and opportunities. Accordingly, a fleet of firms, including catch-up
firms, have been vigorously involved in the process of setting M2M/IoT standards to affect
the direction of technological change. In this context, the investigation of an M2M/IoT
trajectory can be utilized to identify strategic patents and strategic groups, which have a
high potential for strategic alliances. This will help us to understand the dynamics of
strategic competition, revolving around the M2M/IoT trajectory, in the catch-up context.
In our paper, we have the following main research objectives: 1) to demonstrate
how standards affect technological convergence in terms of technological importance and
technological diversity; and 2) to examine whether standards hold a crucial role in the
shaping of the M2M/IoT trajectory and, if so, which standards are significant. Furthermore,
we attempt to identify strategic patents and strategic groups that center around the
M2M/IoT trajectory. To this end, we rely on the dataset of M2M/IoT patents and the
methodologies of social network analysis, main path analysis, patent text similarity analysis,
and clustering analysis. Findings from these analyses are expected to fill the aforementioned
research gaps, contribute to the theoretical understanding of the relationship between
standards and innovation, and offer practical implications.
As for the structure of the paper, the subsequent section contains the review of prior
literature on M2M/IoT as a large technological system, the relationship between
technological trajectories and standards, and the relationship between standards and patents.
Section three describes our research methodology and data collection. We then present the
findings and discussion of this research in sections four and five. Thereafter, the paper is
concluded with the clarification of its findings and limits.
2. Literature Review
2.1. M2M/IoT as a large technological system
The IoT encompasses an ensemble of heterogeneous technologies, standards and subsystems
labeled with reference to three different paradigmatic visions (i.e. things-, Internet-, and
semantic-oriented) (Atzori, Iera, & Morabito, 2010). The system of interconnected artifacts
has been evolving in a way to not only extract information from the environment (sensing)
and interact with the real world (actuation/command/control), but also draw on Internet
standards to offer services for data analytics and applications. The main objectives of this
technological system are apparently to embed intelligence into the environment, enable
2 The full names of IoT technologies and standards-setting organizations and consortia are noted in Appendix 2
(glossary).
Page 5
4
ubiquitous computing without human intervention and vanish the presence of technologies
from the consciousness of people in daily life. To this end, smart connectivity with existing
networks and context-aware computation using network resources are indispensable (Gubbi,
Buyya, Marusic, & Palaniswami, 2013).
Current IoT research predominantly focuses on the technical aspect of IoT,
particularly with respect to system architectures and networks. For instance, Shen & Carug
(2014) argued that the existing architecture reference models (e.g. Open System
Interconnection (OSI) and Next Generation Network (NGN)) are not suitable for IoT,
suggesting a new reference model for IoT standardization. Zhou (2012) introduced a four
pillar of IoT (i.e. M2M, RFID, WSN and SCADA) categorized by different IoT networks.3 Yet
some scholars (e.g. Shin (2014)) viewed IoT as a socio-technical system in which the social
and technical aspects of IoT are intertwined in a complex manner. From this perspective, it is
of significance to investigate the historical development of IoT technologies within the social
context in which they are embedded (Bijker, 1995). As Pinch and Bijker (1987) pointed out,
those technological problems were defined within the context of meaning assigned by
relevant social groups. Hughes (1983), for example, probed into the history of the electric
power system development in Western society, and revealed how artifacts were socially
constructed to become the components of a large technological system.
Standards serve as a medium of coordination among technological artifacts and
social actors (Schmidt & Werle, 1998). Through the processes of social negotiation, consensus
formation and legitimacy seeking, a number of interests of heterogeneous stakeholders are
coordinated and inscribed into standards (de Vries, Verheul, & Willemse, 2003). Standards-
setting committees and their institutional rules offer arenas in which actors negotiate and
contextualize their decision-making processes. Issues regarding compatibility between
components in the system and balance in their performance capabilities have been crucial in
the standards-setting process of emergent network technologies, driven by various social
actors (David & Bunn, 1988). The development of M2M/IoT as a large technological system
particularly necessitates committees-based compatibility standards, which specify the
relational properties of technological artifacts embodied in the system of M2M/IoT patents.
Within a large technological system, the functioning of one subsystem is greatly
affected by its system architecture and complementary subsystems. Technological
imbalances between system elements critically retard a system improvement process
(Rosenberg, 1969). Similarly, Hughes (1987) used the term reverse salients (i.e. “components
in the system that have fallen or are out of phase with others”, p. 73) to describe a set of
critical problems in need of attention for system growth. The sequential patterns of problem-
solving activities arising from interdependencies among components in a large technological
system (i.e. a change in one part becomes a problem in other parts) and the significance of
compatibility naturally lead to the necessity of understanding the path-dependent
development (like trajectories) of technologies and the role of standards.
3 WSN (wireless sensor network) refers to a system of sensors interconnected mostly through the short-range
wireless mesh networks (e.g. Zigbee). While WSN put an emphasis on data gathering from numerous sensor
nodes and unidirectional data transfers to servers via gateways, M2M networks tend to focus on device control
through bidirectional communications between devices and servers (H. Kim, 2014). SCADA (supervisory control
and data acquisition) refers to a smart system, based on closed-loop control theory, which connects equipment
mostly via wired short-range networks (e.g. field buses).
Page 6
5
2.2. Technological change, trajectory and catch-up opportunities
Technological change, propelled by a series of problem-solving activities, is often explained
from the evolutionary perspective of variation, selection and retention (e.g. Nelson & Winter,
1982). The process of variation and selection, in particular, is greatly influenced by a
technological paradigm—“a ‘model’ and a ‘pattern’ of solution of selected technological
problems, based on selected principles derived from natural sciences and on selected material
technologies” (Dosi, 1982, p. 152). An italic emphasis on selected indicates that the context of
choosing relevant problems and appropriate solutions in which a group of people, such as
engineers, are embedded constitutes the core of a technological paradigm. This concept is
linked to Nelson and Winter's (1977) heuristic search processes—“a set of procedures for
identifying, screening, and homing in on promising ways to get to [an activity’s] objective or
close to it. The procedures may be characterized in terms of the employment of proximate
targets, special attention to certain cues and clues, and various rules of thumb” (pp. 52–53).
By delineating the boundaries of perceivable technological possibilities, a technological
paradigm steers the possible directions of technological change.
Due to the cumulative, irreversible nature of technological change, technology
advances along a technological trajectory—“the pattern of ‘normal’ problem solving activity
(i.e. of ‘progress’) on the ground of a technological paradigm” (Dosi, 1982, p. 152). This
pattern emerges through the improvement of multi-dimensional trade-offs among variables
with technical, process, service characteristics (Saviotti & Metcalfe, 1984) in a socially
constructed design space. Given that technology is the embodiment of useful knowledge—a
combination of propositional (how nature works) and prescriptive (how to use techniques)
knowledge (Mokyr, 2002)—with reference to problem-solving activities, a technological
trajectory can be represented by the main flow of knowledge predicated upon a
technological paradigm. Recently, some researchers (e.g. Martinelli, 2012; Verspagen, 2007)
attempted to identity technological trajectories via top paths in patent citation networks,
considering a patent is a useful proxy for engineering knowledge.
Changes in technological systems and techno-economic paradigms affect the cost of
entry in a market and opens catch-up opportunities for de novo entrants (Perez & Soete, 1988).
The paths of technological evolution assume a significant role in catch-up since many
latecomer firms set different types of catch-up strategies grounded on technological
trajectories: path-creating, path-skipping and path-following catch-up (Lee & Lim, 2001).
Standards offer latecomer firms an opportunity to catch up. By consolidating the meaning of
a technical artifact into a common understanding (i.e. “closure” in Bijker's (1995) term),
standards shape technological trajectories by stabilizing the variations of technological
change and channeling them into a certain direction. By developing standards as guideposts
for the evolutionary pattern of problem-solving activities, the tyranny of combinatory
explosion in a multi-dimensional design space will be reined (Metcalfe & Miles, 1994). When
standards, which affect the rate and direction of technological change, are fixed, there is less
risk of choosing which technological trajectory latecomer firms enter and it is easier for them
to produce complementary goods and services (Greenstein & Khanna, 1997; Lee, Lim, &
Song, 2005). Latecomer firms from late-industrialized countries (e.g. South Korea and China)
have strategically utilized standards not only to follow the paths of forerunners, but to
achieve stage-skipping and path-creating catch-ups, particularly in ICT-related industries
Page 7
6
(Lee et al., 2005; Mu & Lee, 2005). This stream of literature highlights the strategic
importance of the M2M/IoT technological trajectory shaped by standards.
2.3 Standards and patents
Depending on coordination mechanisms, standards can be roughly classified into de facto
standards (market) and formal standards (committee) (David & Greenstein, 1990). Formal
standards are established through the processes of discussion, negotiation and consensus-
based agreement by interested parties, while de facto standards are produced via the
process of a dynamic bandwagon in the market without formal agreement or even
discussion (Farrell, 1989). Formal standards, in particular, accentuate the processes of
negotiation among relevant social actors and of consensus with respect to the meaning of
technological artifacts. De facto standards (or dominant designs) put emphasis on
technological dominance in a market (Suarez, 2004).
Standards are often regarded as “public goods” (i.e. non-rivalrous and non-
excludable according to Samuelson's (1954) criteria) (Kindleberger, 1983). Patents, by
contrast, are defined as a set of exclusive rights granted to inventors in return for public
disclosure of inventions. Earlier studies have focused on this inherent tension between
standards and patents. Farrell (1989) explained how the protection of intellectual property,
such as patents, affects the processes of formal and de facto standardization. Strong patent
protection, as he points out, may retard formal standardization, in particular, due to an
increase in vested interest. Accordingly, many standards-setting organizations (SSOs) set
their rules that influence the patenting behaviors of firms by obligating their participatory
members to declare patents that are essential to the implementation of standards (Lemley,
2002).
Recent studies have delved into the value of the (declared) standards-essential
patents (hereinafter “essential patents”). Rysman and Simcoe (2008) found that essential
patents receive more citations over a longer period of time, as compared to non-essential
patents. They also showed that there is a substantial increase in the forward citations of
essential patents after the disclosure to the SSOs, demonstrating the marginal effect of an
SSO endorsement on the value of patents. Bekkers, Bongard, and Nuvolari (2011)
demonstrated that, in comparison with the intrinsic technological value of the patents, a
strategic involvement in the process of standardization serve as a stronger determinant of
essential patents in the case of WCDMA. Similarly, the findings of Kang and Motohashi
(2015) showed that the attendance of inventors to the standardization meetings is a key
determinant of essential patents.
A strong patent portfolio is viewed as one of the key assets to win a standards
competition by bolstering a firm’s position in the standards negotiations (Shapiro & Varian,
1999a). Opportunities to flex their muscles in the negotiations of standards development
may incentivize firms to engage in active patenting activities (Gandal, Gantman, &
Genesove, 2007). For instance, leveraging its large portfolio of essential patents, Motorola
was able to exert a formidable negotiation power in the process of the GSM standardization
(Bekkers, Verspagen, & Smits, 2002). Intriguingly, Blind and Thumm (2004) found that firms
with higher patent intensities tend not to join the process of formal standardization,
indicating that technological advantages based on strong patent portfolios may guide firms
to pursue de facto standardization. That is to say, a great patent portfolio gives a firm the
Page 8
7
upper hand in a knowledge position, empowering it to make various strategic moves. With
the respect to knowledge positions in standards-based high-tech markets, Bekkers and
Martinelli (2012) showed that link connectivity based on the measure of traversal counts in
the top main paths is a better indicator than the simple counts of essential patent families.
These earlier studies lend support to the investigation of an M2M/IoT technological
trajectory as a means to understand firms’ technology strategies.
To the best of our knowledge, there are very few studies that have investigated
citations to non-patent literature (NPL) as a linkage between standards and patents, while
some past research has discussed the role of science in patents through the examination of
references to NPL. Citations to NPL have been used as an indicator to operationalize the
dependence of technology on science. As Tijssen (2001, p. 52) remarked, “these citations at
the very least indicate an awareness of scientific results with some indirect bearing on
elements of the invention. In the best case, they reflect strong evidence of substantial direct
contributions of scientific inputs to breakthrough technological innovations.” Fleming and
Sorenson (2004) found that references to science can be of great use when inventors give an
attempt to solve difficult problems involving the combination of highly coupled (i.e. non-
modular) components. These science linkages work as part of the context in which
technological innovations are to be situated (Narin, Hamilton, & Olivastro, 1997). This logic
underneath NPL as a science linkage to patents is similarly applicable to the relationship
between standards and patents. Rysman and Simcoe (2008), for instance, regarded patents
referencing to IETF standards in their non-patent literature as reflecting platform
technologies in that those patents are more likely to build upon the Internet standards.
Assumed that standards influence the shaping of a technological trajectory, which
can be identified by a main path in the patent citation network, patents citing standards are
more likely to serve as a base for subsequent technological changes. The basicness of patents
can be operationalized by two indicators: importance and generality (Trajtenberg,
Henderson, & Jaffe, 2002). The importance can be measured by the number of forward
citations, while the generality can be calculated by the Herfindahl index on technological
classes of citing patents.4 Within the field of network technologies where compatibility
assumes a critical role (David & Steinmueller, 1994), patents referencing to standards that
ensure compatibility are likely to be grounded upon core knowledge that is fundamental to
interoperability within a large technological system, and thereby receive more citations.
Accordingly, the following hypothesis can be derived:
H1: a reference to standards is positively associated with the importance of patents.
Standards offer a function of variety-reduction that assists industry players in mitigating
risks and achieving focus, which is critical to scale economies-based market expansion
(Blind, 2004). By optimizing the variety of technologies and selecting dominant designs,
standardization works as the process of freezing technological frames and ushering a
technology cycle into the era of incremental change (Anderson & Tushman, 1990). Standards,
4 According to Trajtenberg et al. (2002), generality refers to the extent to which subsequent technologies are
spread across different fields, rather than being concentrated in just a few of them.
GENERALi = 1 – ∑ (NCITING𝑖𝑘
NCITING𝑖)2𝑁𝑖
𝑘=1 where k is the index of patent classes (IPC), and Ni is the number of different
classes to which the citing patents belong.
Page 9
8
selected by industry players as a part of technological infrastructure, generate self-
reinforcing mechanisms and lock-in effects, and thereby may hamper flexibility (Arthur,
1994; Hanseth, Monteiro, & Hatling, 1996). This means that standards-citing patents are
more likely to have lower generality values, which represent higher concentration in a
specific technological field. Hence, we drew the following hypothesis:
H2: a reference to standards is negatively associated with the generality of patents.
Patent citation networks have been used to investigate the evolutionary patterns of
technological convergence. The convergence of heterogeneous technologies often propels
the driving forces of technological change. Technological convergence is defined as “the
process by which two hitherto different industrial sectors come to share a common
knowledge and technological base” (Athreye & Keeble, 2000, p. 228). For instance, common
processes were shared among different machinery sectors (Rosenberg, 1976), and certain
generic technologies were imperatively used as an essential base for product innovation in a
number of the world’s largest firms (Patel & Pavitt, 1997). The convergence of information
and communications technologies (ICT) has been driven by sharing the standards of
common network architectures. GSM (2G) and UMTS (3G) standards, for example, assumed
a critical role in the emergence of new technological trajectories in the mobile telecom
industry, generating market potential for strategic alliances to add new services compatible
with the shared network standard (Sadowski, Dittrich, & Duysters, 2003).
The similar structures of convergence in the spheres of patents and standards
indicate a correlation between technological development and standardization (Gauch &
Blind, 2015). Patents related to services adapted for wireless communication networks (IPC
H04W4/00), in particular, among essential patents have held a central position in
technological convergence in ICT standards (Han & Sohn, 2016). The degree of attraction
and that of technological diversity have been employed to measure technological
convergence in patent citation networks (Cho & Kim, 2014). The importance of patents and
the generality of patents correspond to the degree of attraction and that of technological
diversity, respectively. Since standards serve as a base of technological convergence, the
following hypothesis can be derived:
H3: a reference to standards moderates the relationship between the importance of
patents and the generality of patents.
3. Methodology
3.1. Patent citation-based communities and main path analysis
Patents have been long recognized as a rich data repository for the study of technological
change. Patent citation data (backward citations refer to the number of cited documents and
forward citations refers to that of citing documents) enable epistemic linkages to be traced in
a technological field. The premise underlying patent citation analysis is that highly cited
patents may contain essential technological advances which later technologies are built upon
(Karki & Krishnam, 1997). This possibility is predicated on a cumulative and irreversible
view of technological progress, by which inventions are influenced by prior arts and in turn
contribute to the knowledge base for future inventions (Jaffe & Trajtenberg, 2002). In this
sense, the network of patent citation data may reflect the dynamics of technological
Page 10
9
evolution at a knowledge level. Particularly in industries where technological systems are
deeply interdependent, the network analysis of patent citations can be of great use to
discover core and emerging technologies, which is vital for technology strategy formulation
(Cho & Shih, 2011).
The techniques and software packages of social network analysis are frequently
used to examine and visualize patent citation networks. The social network analysis
conceptualizes individual patents as nodes and their citation relations as lines, and reveals
the structural properties of a patent citation network via several measurement indicators (e.g.
centrality) and the graphical representation of the network (Wasserman & Faust, 1994). In
addition, the modularity optimization method (Blondel, Guillaume, Lambiotte, & Lefebvre,
2008) was employed to identity communities in the patent citation network. This method
finds optimal values by measuring the density of links inside communities as compared to
links between communities. Among centrality indicators, betweenness centrality was
primarily used to identify the main flow of knowledge in a technological field. It measures
the number of shortest paths from all nodes to others that pass through a particular node
(Freeman, 1977). Assumed that the transfer of knowledge more frequently occurs along the
shortest paths, the citation network of patents with high betweenness centrality may map
the bedrock of an evolving technological field within an epistemic context. As with previous
research (Leydesdorff, de Moya-Anegón, & Guerrero-Bote, 2015), we used Gephi to
visualize the citation network of M2M/IoT patents and calculate network indicators,
including betweenness and closeness centrality (Freeman, 1979) and modularity (Blondel et
al., 2008).
For the crystallization of the M2M/IoT trajectory, one of the Hummon and Dereian's
(1989) connectivity measures (i.e. Search Path Node Pair (SPNP)) was employed to identify
the top path in the M2M/IoT patent citation network. This connectivity algorithm calculates
a weight value assigned to each citation link based on its relative position in the overall
network structure. If the SPNP value of a particular link is 21, for instance, it means that link
connects a total of 21 pairs of nodes.5 The top path can be identified by tracing the nodes
with the highest SPNP value from sources (i.e. starting points—patents that do not cite any
other patents) to sinks (i.e. end points—patents that do not receive any citations). In several
research papers (e.g. Fontana, Nuvolari, & Verspagen, 2009; Verspagen, 2007), this path is
considered the critical backbone of knowledge flow, representing a main technological
trajectory.
As with prior studies (Fontana et al., 2009; Martinelli, 2012), we used Pajek to
calculate the SPNP values and identify the top path of the M2M/IoT patent citation network.
The main research objectives will be addressed by carefully examining whether patents
lying on the main path of the M2M/IoT trajectory.
3.2. Patent similarity-based clustering analysis
Patents reflect firms’ technological capabilities to innovate. Firms with similar technological
capabilities often take form as a strategic group and similarly act in the face of uncertainty
(Nohria & Garcia-Pont, 1991). Yet patent text similarity raises the risk of potential patent
5 For the details of the calculation method of SPNP, see Fontana, Nuvolari, & Verspagen (2009).
Page 11
10
infringement lawsuits, intensifying competition between firms with similar patents
(Bergmann et al., 2008). That is to say, firms with similar patents are likely to interact in a
simultaneously cooperative and competitive (coopetitive) manner (Brandenburger &
Nalebuff, 1996).
Latent semantic analysis (LSA), based on singular value decomposition (SVD), has
often been used to examine patent text similarity (Han & Sohn, 2015). LSA is a mathematical
computation-based method for extracting the aggregate of word contexts, which determines
the similarity of words (Landauer, Folt, & Laham, 1998). Lee and his colleagues’ research
demonstrated that the performance of the LSA model is consistent with human judgments
of similarity (Lee, Pincombe, & Welsh, 2005). We employed LSA to derive patent text
similarity estimates and identify strategic groups of firms that center around the M2M/IoT
trajectory.
We extracted terms from the title, abstract and main claim of patents,6 and then
preprocessed the text data by first transforming the terms to lower-case letters, second
removing numbers, punctuation, words of length less than three and stop-words, and third
stemming the terms. Thereafter, a term-document matrix was generated based on the
occurrences of terms in patents. The weighting of terms is applied to dampen the effects of
frequent terms in each patent and amplify the effects of infrequent terms across the patents.
Log frequency weighting is employed for local weighting, whereas entropy term weighting
is used for global weighting.
Using R, the software commonly used for statistical analysis, we created a latent
semantic vector space from the term-document matrix. To normalize ratio-scaled values, we
divided each value of the column by the square root of the column sum of squares of all
values. Cosine similarity was used to calculate the similarity of patents. Thereafter, Ward’s
hierarchical agglomerative clustering method was applied to identify patent clusters based
on patent text similarity. Ward’s minimum variance criterion was employed to detect
clusters with minimal within-cluster variances (Ward, 1963). We considered these clusters as
strategic groups in a coopetitve manner.
3.3. Variables
The influence of formal standards on the shaping of a technological trajectory can also be
demonstrated by testing the aforementioned hypotheses (H1, H2, & H3). The technological
trajectory is operationalized by a network of patent citations, and the linkage between
formal standards and patents is measured by a patent citation to standards in non-patent
literature. Therefore, a binary variable indicating whether patents reference to standards
(Standards) is the main independent variable (IV) in this research. For H1, the importance of
patents is measured by the number of forward citations (Trajtenberg et al., 2002). Carpenter,
Narin, & Woolf (1981) demonstrated that technologically important patents underlying
products with industrial awards are more highly cited than a control group of randomly
selected patents. Similarly, the findings of Harhoff, Scherer, & Vopel's (2003) research
showed that the count of citations received from subsequent patents is positively related to
the patent’s value, estimated from a survey of patent-holders. Older patents tend to receive
6 As for the text of standards, we used terms from the title, scope, and overview of standards.
Page 12
11
more citations than newly published patents. To prevent a bias arising from the age effect,
we divide the number of forward citations by the average number of forward citations in the
same application year (Hall, Jaffe, & Trajtenberg, 2001).
For control variables, the number of backward citations (Backward citations) was
first taken into account. This may reflect the dependence on prior technologies and the
cumulativeness of a patent (Cassiman, Veugelers, & Zuniga, 2008). Previous research (e.g.
Harhoff et al., 2003) showed that backward citations are positively associated with a patent’s
value, which is also correlated with forward citations. It is because the more a patent is
within the local search domain of other firms via direct connections to other patents (i.e.
backward citations), the more likely the patent to serve as a base for boundedly rational
firms’ search for future technology (Podolny & Stuart, 1995). For firm-specific variables, we
added the number of employees, the number of patents granted in USPTO and R&D
expenditures as proxies for firm size and general technological assets (LnFirmSize, LnIP,
LnR&D) (Bekkers et al., 2011; Kang, Huo, & Motohashi, 2014; Leiponen, 2008). Larger-sized
firms with greater technological potential have favorable conditions in making more
valuable innovations (Schumpeter, 1950). Those numbers were rescaled by taking the
natural logarithm. Whether firms are founded in catch-up economies was also factored in as
a dummy variable (Catch-up). European and US firms have been holding strong knowledge
positions, which are crucial to produce technologically important patents, particularly in the
field of mobile communications (Bekkers & Martinelli, 2012). Firms in latecomer countries
(such as Korea and China) have traditionally relied on external knowledge from European
and US firms, yet have recently employed technology standards to catch up (Kang et al.,
2014). Application years were inserted as dummy variables to capture any residual temporal
effects (Year).
For H2, the generality of patents is measured by the Herfindahl index on
technological classes of citing patents (Generality) (Trajtenberg et al., 2002). This variable
indicates how much a patent is cited by subsequent patents from various technological fields,
reflecting technological diversity. Its value ranges from 0 (extremely specific) to 1 (extremely
general). For control variables, originality was taken into consideration. Originality is the
backward citation measure, equivalent for generality, indicating the broadness of a patent’s
technological root (Originality). If a patent cites prior research from a wider range of
technological fields, its originality gets close to one. Previous studies (e.g. Trajtenberg et al.,
2002) show that originality breeds generality. The number of forward citations was also
factored in (Forward citations). Where a patent receives more subsequent citations, there is a
built-in tendency to cover more technological classes (Hall et al., 2001). A catch-up firm
dummy variable is inserted as well. Given R&D resource constraints, firms in catch-up
economies tend to strategically select and concentrate on a specific technological field (Park
& Lee, 2006).
The Tobit model was mainly used to test the aforementioned hypotheses (H1: the
number of adjusted forward citations as a dependent variable (DV) (Kang et al., 2014); H2:
the generality of a patent as DV (Cassiman et al., 2008)). This model is often employed to
describe the relationship between IVs and a non-negative DV, in which the observations
take rational numbers and are bounded below by zero. The Tobit model assumes the
homoscedasticity and normal distribution of residuals. Powell's (1984) method for censored
least absolute deviations (CLAD) is often used to estimate the effect of IVs on the conditional
median of DV in the case of the violation of those assumptions, since this non-parametric
Page 13
12
regression method is robust to such violation (Sampat & Ziedonis, 2004). We also performed
robust checks, relying on the Powell’s CLAD estimator with 200 bootstrap replications. For
H1, the number of (non-adjusted) forward citations is additionally considered as DV. The
Poisson regression model is first considered for count data with non-negative integer values
where being cited is an event with a probability of success (Cameron & Trivedi, 1998). Yet in
the case of violation of the equipdispersion assumption (i.e. the variance is equal to the
mean), the negative binomial (NB) regression model is taken into account (Hausman, Hall,
& Griliches, 1984). After the overdispersion test, we drew on the NB model for the number
of forward citations.
For H3, the importance of patents is modified to measure the binding force of a
technology, a source of attraction for technological interaction (Binding force). Closeness
centrality is used as a proxy variable for the binding force (gravity concept), which explains
how each node is cohesively connected to the others (Cho & Kim, 2014). Following the
gravity model, the importance of a patent (the count of adjusted forward citations) was
weighted with the square of the patent’s closeness centrality. A high binding force indicates
a patent with high degree centrality and high closeness centrality. A strong correlation
between binding force (source of attraction) and generality (technological diversity) results
from technological convergence.
Subgroup analysis was performed to probe into the moderating role of standards
(Sharma, Durand, & Gur-Arie, 1981) in the relationship between the binding force and
generality. The dataset was split into two subgroups: a group of patents without a reference
to standards (STD=0) and that with (STD=1). A moderation test was conducted to find
statistical differences between the two groups in the regression coefficients of generality on
binding force from the Tobit model. As with previous research (Cameron & Trivedi, 1998),
we employed Stata to conduct a statistical analysis of count data. We also used R to cross-
checked the test results.
3.4. Data
For data collection, US patent data was extracted from WIPS, one of the largest online
subscription-based patent databases in South Korea. We double-checked the accuracy of
patent data by matching it with data from the database of USPTO (United States Patent and
Trademark Office). US patent data has been widely used to investigate the dynamics of
innovation under the assumption that it reflects technological developments worldwide
(Erdi et al., 2013). US patent citations are particularly valuable due to the duty of candor
under the US patent law, which includes that a failure to disclose relevant prior art may
render any ensuing patent unenforceable (Erstling, 2011; Gay, Le Bas, Patel, & Touach, 2005).
The Cooperative Patent Classification (CPC) code7 H04W4/005 (mobile application
services or facilities specially adapted for wireless communication networks for machine-to-
machine communication [M2M, MTC]), was used to set the patent search scope. The number
of granted M2M/IoT patent has rapidly increased in recent years. In 2014, there were 235
granted M2M/IoT patents under the CPC code H04W4/005. Yet as of November 16, 2016, a
7 CPC, as an extension of the International Patent Classification (IPC) code, has been jointly developed by the
USPTO and EPO (European Patent Office).
Page 14
13
total of 1203 granted M2M/IoT patents were searched under the same code. This makes
evident that M2M/IoT is a newly emerging, fast-growing technology field. In this paper, we
used 588 M2M/IoT patents, which were granted before the year 2016, and extracted their
backward/forward citation data. In total, 5,425 patent citation data was used to visualize and
analyze the M2M/IoT patent citation networks.
The dataset of granted patents is subject to the truncation issue (i.e. missing
observations of patents filed in recent years that have not yet been granted) (Hall et al., 2001),
as shown in Figure 1. In particular, the forward citations of recently applied patents may not
be fully observed as a result of truncation. To minimize the problem of truncation, we select
five years’ patents with application years, ranging from 2008 to 2012, for the statistical
analysis. A past five year window has been used to measure the current impact index (Karki
& Krishnam, 1997). The application year 2008 is the year when formal standards with
respect to M2M/IoT were first cited in the M2M/IoT patents. From Compustat via WRDS8
and the USPTO database, we retrieved data on the average number of employees, the
average R&D spending and the average number of patents granted in USPTO over the
period 2008–12 for the firms included in our dataset. The descriptive statistics of the patent
data over the period 2008–12 is presented in Table 1. The correlation matrix is shown in
Table 2.
Figure 1. Number of granted M2M/IoT patents
Table 1. Descriptive statistics of the patent data over the period 2008–129
Variable Description Obs. Mean SD Min. Max.
Standards Dummy on whether a citation is made
by a patent to a formal standard
421 0.49 0.50 0 1
Forward citations Count of citations that a patent receives
from subsequent US patents
421 2.93 5.86 0 56
Adjusted forward Number of forward citations divided 421 1.00 2.06 0 28.65
8 https://wrds-web.wharton.upenn.edu/wrds/
9 The IPC classes of foreign patents that cited, and were cited by, M2M/IoT patents were included to measure
generality and originality.
Page 15
14
citations by the average number of forward
citations in the same application year
Generality Herfindhal index (IPC classes) on
forward citations
253 0.44 0.25 0 0.91
Originality Herfindhal index (IPC classes) on
backward citations
253 0.65 0.20 0 0.93
Binding force Count of adjusted forward citations
weighted with the square of closeness
centrality
376 0.71 1.74 0 27.65
Backward
citations
Count of citations made by a patent to
prior US patents
421 20.62 70.88 0 1209
Catch-up Dummy on whether firms are from
catch-up economies (i.e. Korea, China
and Taiwan)
421 0.31 0.46 0 1
LnFirmSize Natural log of average employees 394 10.25 2.33 0.69 12.94
LnIP Natural log of average US patents 414 7.01 2.39 0 10.17
LnR&D Natural log of average R&D
expenditures
344 21.46 1.69 14.82 24.21
Table 2. Correlation matrix
STD FCITE GEN ORI BIND BCITE CAT SIZE IP
STD
FCITE –0.05
GEN –0.22 0.27
ORI –0.23 0.18 0.37
BIND –0.04 0.23 0.20 0.10
BCITE –0.14 0.52 0.21 0.16 0.75
CAT 0.19 –0.09 –0.35 –0.28 –0.11 –0.11
SIZE 0.13 –0.15 –0.14 –0.09 –0.09 –0.19 0.17
IP 0.21 –0.03 –0.23 –0.11 –0.03 –0.08 0.32 0.64
RD 0.02 –0.16 0.11 –0.01 –0.07 –0.05 –0.11 0.72 0.69
STD: Standards, FCITE: Forward citations, GEN: Generality, ORI: Originality, BIND: Binding force,
BCITE: Backward citations, CAT: Catch-up, SIZE: LnFirmSize, IP: LnIP, RD: LnR&D.
4. Findings
4.1. M2M/IoT communities and strategic patents
Drawing upon the Gephi software, we visualized the overall M2M/IoT patent citation
network structure, as shown in Figure 2 (Left). This figure shows that there is a giant
component, consisting of several clusters, in the center of the M2M/IoT patent citation
network. It is obvious that some patents serve as strategic linkages that connect different
clusters in the giant component. By relying on modularity optimization (Blondel, Guillaume,
Lambiotte, & Lefebvre, 2008),10 we identified ten different communities in the giant
component of the M2M/IoT patent citation network, as shown in Figure 2 (Right).
10 We set ten as a resolution limit on the size of the smallest community.
Page 16
15
Modularity measures the fraction of edges within communities (i.e. clusters or
modules) as compared to the expected fraction of such edges placed at random.11 It is used
to identify densely connected groups of nodes with sparse edges between groups, which
reflects community structure in a network. A good partition of communities refers not
merely to one in which there are few edges between communities, but to one in which there
are fewer than expected edges between communities (Newman, 2006). Using modularity as
a goodness measure of community structure is one of the most popular community
detection methods (Fortunato, 2010). Blondel et al.’s algorithm, first, optimizes modularity
by allowing only local changes of communities over all nodes, and then aggregates the
found communities in order to find a global maximum of modularity (Blondel et al., 2008).
The modularity of the top ten communities in the giant component of the M2M/IoT
networks is above 0.8. Generally, above 0.3 is considered a good indicator of significant
community structure in a network (Clauset, Newman, & Moore, 2004). The significant
structure of the top ten communities can also be confirmed by the visualization of the
densely connected groups of nodes with sparse edges between communities in the M2M/IoT
network, as manifested in Figure 2. This means that patents in the same community share
the same backward citations or forward citations, and thereby contain similar information.
Therefore, each community corresponds to a specific type of technology that is distinct from
one another. The emergence of this technology community takes place since firms with
similar strategic interest cite similar standards or patents. These communities can also be
considered as strategic blocks, which are based on similar linkages (Nohria & Garcia-Pont,
1991).
For the verification of the community detection results, we used a different
algorithm (i.e. Chinese Whispers) to check whether similar community structure is
identified in the network. The Chinese Whispers (CW) algorithm is used to find groups of
nodes that deliver the same information to their neighbors (Biemann, 2006). Similar with
Blondel et al.’s modularity optimization method, the CW algorithm works in a bottom-up
fashion. Each node inherits the strongest class (or information) in its neighborhood (i.e. the
class whose sum of edge weights to the given node is maximal). A group of the same class
(i.e. cluster) stabilizes during the iteration, and its boundary expands until it meets the
border of another cluster. By employing the CW algorithm, we found the cluster structure
that is similar with the one identified via Blondel et al.’s approach, as shown in Appendix 1.
The only noticeable difference is that the CW algorithm could not group together nodes
within the first community identified by Blondel et al.’s method. This implies that the first
community contains relatively inhomogeneous information as compared to other
communities.
11 It is defined as Q =
1
2𝑚∑ [𝐴𝑖𝑗𝑖,𝑗 −
𝑘𝑖𝑘𝑗
2𝑚]𝛿(𝑐𝑖 , 𝑐𝑗), where 𝐴𝑖𝑗 represents an element of an adjacency matrix whose
value is 1 if nodes i and j are connected and 0 otherwise, 𝑘𝑖(= ∑ 𝐴𝑖𝑗𝑗 ) is the sum of the edges attached to node i,
𝑐𝑖 is the community to which node i is assigned, the 𝛿-function 𝛿(𝑢, 𝑣) is 1 if u = v and 0 otherwise, and m (= 1
2∑ 𝐴𝑖𝑗𝑖,𝑗 ) is the total number of edges in a network.
𝑘𝑖𝑘𝑗
2𝑚 is the probability of an edge existing between nodes i
and j if connections are made at random but respecting node degrees (Clauset et al., 2004).
Page 17
16
Figure 2. The overall citation network structure of M2M/IoT patents (Left) and top 10 communities (Right)
The community detection method is of great use to discover nodes which hold a central
position within each community and those which serve as bridges between communities.
Relying on betweenness centrality, we extracted the top ten patents with a central position
within each community in the M2M/IoT patent citation, as shown in Table 3. In the first
community, which is the largest cluster and colored green, several different firms are
connected, and, in turn, there is no single dominant firm. In the second community, colored
sky blue, Seven Networks and Intel are main players. In 2013, Seven Networks announced
its plan to collaborate with Intel in order to help operators manage app data traffic and
optimize their networks (Hill, 2013). By looking at this collaboration and the titles of top ten
patents, we believe that the second community is related to the technology of software-
defined networking (SDN).
The third and seventh communities are inter alia notable. The third community,
which is colored pink, is unambiguously dominated by Google and Nest Labs. It appears
that Google has been building up its own knowledge pool, which is recognized by the
pattern of self-citations. Self-citation is viewed as reflecting the appropriation of returns (i.e.
the more frequently subsequent inventions that occur “in-house”, the greater would be the
benefits reaped by the original inventor (Trajtenberg et al., 2002)). After the acquisition of
Nest Labs in Jan 2014, its second-largest bid ($3.2 billion) (The Economist, 2014), Google has
been putting its innovation efforts into programmable thermostats as a controlling device for
the system of connected home appliances. A series of self-cited Google and Nest patents
(8478447, 8752771, 8757507, 9026254, 9092040, 9104211, 9223323) primarily focus on the
development of user-friendly, multi-sensing, self-recharging, and learning thermostats as a
control unit for heating ventilation and air-conditioning (HVAC) systems. Similarly,
Amazon has been generating its own M2M/IoT technological path in the seventh
community, colored light green. The seventh community is composed mainly of Benjamin’s
patents, including a few of older M2M/IoT patents (7058356, filed in 2001, and 7496328, filed
in 2005). Amazon obtained the ownership of subsequent Benjamin’s patents (8620208,
8666308, 8718539) and localized the stock of this accumulated knowledge. This set of patents
Page 18
17
center mainly around interactions between hand-held devices (e.g. mobile phones) and
remotely located entities with payment modules (e.g. TV).
Table 3. The assignees of the top 10 patents in the top 10 communities in the M2M/IoT patent citation network
1
(26.36%)
2
(14.05%)
3
(8.46%) 4
(7.3%) 5
(6.21%)
6
(5.95%) 7
(5.2%)
8
(3.19%)
9
(2.95%)
10
(1.73%)
InterDigita
l (8718688)
Seven
Network
s
(9060032)
Google
(9223323)
Mueller
Intl
(8855569)
Robin Dua
(8244179)
iRobot
(8265793)
Benjamin
Slotznick
(7058356)
Nielsen
(8930003)
CNH
America
(8280595)
IBM
(7769848
)
Symstrea
m
(8135362)
Seven
Network
s
(9173128)
Google
(9026254)
Magee
Scientific
(6317639)
Robin Dua
(8583044)
DexCom
(9028410)
Benjamin
Slotznick
(7496328)
Qualcom
m
(8606293)
SynapSens
e (7995467)
Infineon
(8787266
)
Apple
(8737989)
LG
(8811961)
Google
(9092040)
Rain Bird
(8649907)
Robin Dua
(8463184)
iRobot
(8892260)
Qualcom
m
(8831568)
Digimarc
(8417793)
Blueforce
(8467779)
N/A
(8041772
)
Ericsson
(8407769)
Intel
(8619654)
Google
(9104211)
Rain Bird
(8849461)
Robin Dua
(9020429)
Welch
(8458149)
Benjamin
Slotznick
(8131208)
N/A
(8094949)
N/A
(7917167)
N/A
(8498224
)
Via
Telecom
(8737265)
AT&T
(9043503)
N/A
(8489243)
N/A
(6954701)
Robin Dua
(8548381)
iRobot
(8577501)
N/A
(8472935)
ETRI
(8116243)
IBM
(7822852)
Verizon
(8345546
)
Samsung
(9084074)
HTC
(8438278)
Nest Labs
(8752771)
N/A
(6560543)
Robin Dua
(8768256)
DexCom
(8844007)
Benjamin
Slotznick
(7856204)
Digimarc
(9084098)
IBM
(8898284)
Sprint
(8848558
)
Qualcom
m
(9210527)
Intel
(8818376)
Nest Labs
(8757507)
Sun
Micro
(7130773)
Robin Dua
(8971803)
DexCom
(9002390)
Qualcom
m
(8868038)
Digimarc
(8671165)
SkyBitz
(8971227)
Sprint
(9160629
)
Huawei
(9107226)
Intel
(9130688)
Nest Labs
(8478447)
N/A
(7424399)
Robin Dua
(9160419)
N/A
(7639715)
Amazon
(8620208)
Nat’l
Taiwan
Univ
(7839764)
Blueforce
(8682309)
Sprint
(9154976
)
Cellco
(7774008)
Alcatel
Lucent
(9208123)
Samsung
(8392597)
Harris
(8314717)
Robin Dua
(9160420)
N/A
(7327732)
Amazon
(8718539)
Digimarc
(8943172)
Blueforce
(9066211)
Cubic
(8824444
)
Alcatel
Lucent
(9071925)
Huawei
(9173244)
Ericsson
(9049104)
N/A
(7890568)
Robin Dua
(9014631)
N/A
(6456875)
Amazon
(8666308)
Tangoe
(9191523)
N/A
(8233463)
Cubic
(8824445
)
Note: In parentheses are the percentage of patents, measuring the size of a community, and patent numbers.
Relying on the concept of “structural holes”(the separation between non-redundant contacts)
(Burt, 1992), we identified strategic patents of M2M/IoT, which enabled brokering of the
knowledge flow between different patent communities. In Burt’s explanations, the lack of
ties between different groups generates disparities in information held by the groups, and
accordingly a tie that links previously disconnected networks offers strategic benefits of
access to non-redundant information. The brokering of a technology introduces new
knowledge by different combinations of existing knowledge from disparate communities
(Hargadon & Sutton, 1997).
To identify strategic patents, we first extracted patents with high betweenness
centrality (up to top 5%). Nodes with high betweenness centrality are of significance since
they function as points for control of information flow in the network (Freeman, 1977).
Thereafter, we identified strategic patents that link different communities, as shown in Table
4. Considering that the transfer of non-redundant information is more likely to take place
along the strategic linkages between different patent clusters, firms which own strategic
Page 19
18
patents may have control over a source of future M2M/IoT innovation. For instance, Via
Telecom’s patent 8737265 influences the first community (the largest M2M/IoT cluster) by
connecting it with the second community (related to traffic management and network
optimization). Qualcomm’s patents 8831568 and 8868038 affect the seventh community
(connection between M2M devices and mobile communication networks) by connecting it
with the first community.
Table 4. Strategic patents of M2M/IoT
Assignee (Number) Application
Year
Title Connected
Communities
(In/Out)
Via Telecom
(8737265)
2011 Methods and apparatuses for machine type communication 1 <- 2 (In)
LG (8811961) 2011 Method and apparatus for MTC in a wireless communication
system
2 -> 1 (Out)
Seven Networks
(8438633)
2006 Flexible real-time inbox access 1 -> 2 (Out),
1 <- 2 (In)
Seven Networks
(9173128)
2013 Radio-awareness of mobile device for sending server-side control
signals using a wireless network optimized transport protocol
2 <- 1 (In)
Seven Networks
(9060032)
2012 Selective data compression by a distributed traffic management
system to reduce mobile data traffic and signaling traffic
2 -> 1 (Out)
Vodafone (8838806) 2011 Connection management for M2M device in a mobile
communication network
1 -> 7 (Out)
Qualcomm
(8868038)
2012 Methods of and systems for remotely configuring a wireless device 7 <- 1 (In)
Qualcomm
(8831568)
2012 Automatic configuration of a wireless device 7 <- 1 (In)
Magee Scientific
(6317639)
1999 Automatic wireless data reporting system and method 4 -> 8 (Out)
Digimarc (8094949) 2000 Music methods and systems 8 <- 4 (In)
Benjamin Slotznick
(7058356)
2001 Telephone device with enhanced audio-visual features for
interacting with nearby displays and display screens
7 -> 9 (Out)
Iwao Fujisaki
(7917167)
2008 Communication device 7 <- 9 (In)
4.2. M2M/IoT trajectory shaped by formal standards (3GPP in particular)
Using the metric of betweenness centrality (top 5%), we visualized the citation network of
influential M2M/IoT patents, as shown in Figure 3 (Left). Strategic patents that link different
communities are displayed in the figure. This visualization also demonstrates that the first
community, colored green, holds a crucial role in the M2M/IoT network. Relying upon
Hummon and Dereian's (1989) SPNP calculation method, we identified the top path of the
M2M/IoT patent citation network. This method can identify edges with high betweenness
centrality. Recently, several researchers relied on this method to find technological
trajectories, which represent technological innovations as sequential and interrelated events
(Barberá-Tomás, Jiménez-Sáez, & Castelló-Molina, 2011; Fontana et al., 2009; Verspagen,
2007). It assumes that there is a high degree of technology cumulativeness along the
trajectories.
The technological trajectory of M2M/IoT, which penetrates the center of the first
community, is shown in Figure 3 (Right). Table 5 shows nine patents that center on the
M2M/IoT trajectory. Six trajectory patents are also among the top ten patents with high
Page 20
19
betweenness centrality in the first community, as shown in Table 3. This supports the
robustness of this finding. Those six patents are InterDigital (8718688), Apple (8737989),
Ericsson (8407769), Qualcomm (9210527), Huawei (9107226),12 and Alcatel Lucent (9071925).
Figure 3. Patents with high betweenness centrality (top 5%) (Left) and M2M/IoT trajectory (Right)
Table 5. Patents that center on the M2M/IoT trajectory
Assignee
(Number)
Application
Year
Title Standards
Ericsson
(8407769)
2008 Methods and apparatuses for machine type
communication
3GPP TR 33.812, 3GPP TS
43.020
Apple (8737989) 2008 Methods and apparatus for machine-to-machine based
communication service classes
3GPP TS 23.008, ITU-T
Q.763
InterDigital
(8718688)
2010 Method and apparatus for solving limited addressing
space in machine-to-machine (M2M) environments
3GPP TR 22.868, 3GPP TS
23.003, 3GPP TS 23.060
Qualcomm
(9210527)
2011 Method and apparatus for providing uniform machine-
to-machine addressing
Alcatel Lucent
(9071925)
2011 System and method for communicating data between an
application server and an M2M device
3GPP TS 22.368
HTC (9167470) 2011 Handling signaling congestion and related
communication device
3GPP TS 22.368. 3GPP TS
23.107, 3GPP TR 23.888,
IEEE Std 802.11
LG (9137624) 2012 Method and device for performing ranging in a wireless
communication system
Huawei
(9107226)
2012 Method and system for handling congestion in a
communications system
IEEE P802.16b/D2, IEEE
P802.16p/D3, IEEE 802.16m-
08/004r2
Sprint (8638724) 2012 Machine-to-machine traffic indicator
Intriguingly, the majority of trajectory patents (6 out of 9) referenced to formal standards, as
shown in Table 5. In addition to the M2M/IoT trajectory, the first community is composed
primarily of patents that cited formal standards, as shown in Figure 4 (Left). Among the
12 The assignee of this patent is Futurewei, which is Huawei’s subsidiary located in the US.
Page 21
20
patents with high betweenness centrality (top 5%) in the first community, more than half of
the patents (83 out of 141) cited formal standards. This demonstrates that formal standards
affect the formation of the main cluster and the trajectory in the M2M/IoT patent network.
Particularly, 3GPP standards exert a significant influence. Among the formal standards that
were cited by M2M/IoT patents, the proportion of 3GPP is dominantly higher than that of
other standards, as shown in Figure 4 (Right). Among the high betweenness centrality
patents that cited formal standards in the first community, approximately 80% (66 out of 83)
of the patents referenced to 3GPP standards.13 3GPP TR 22.868, TR 33.812, TS 22.368, TR
23.888, in particular, are referenced by many M2M/IoT patents in the standards-shaped
patent cluster. These standards are drafted by 3GPP technical specification group, called
Service and System Aspects (SA). This group is responsible for the overall architecture and
service capabilities of M2M/IoT systems. This indicates that architectural knowledge
embodied in 3GPP SA standards set the boundaries of combinatory explosion in the
M2M/IoT technological evolution and guided multiple variations in a certain direction.
Figure 4. Standards-referencing patents in the citation networks with high betweennness centrality (top 5%).
Note: In the left, standards-referencing patents are colored dark red. In the right, patents that reference to
different standards are differently colored (3GPP (Blue), IEEE (Red), ETSI (Green), IETF (Yellow), mixed/others
(Grey)).
4.3. Technology clusters along the M2M/IoT trajectory
The first community is based on relatively inhomogeneous technologies as compared to
other communities, as shown in Figure 2 and Appendix 1. Yet Blondel’s et al.’ algorithm
cannot detect smaller clusters within the first community, as it focuses on the sharing of the
same edges. In order to find smaller clusters along the M2M/IoT trajectory which penetrates
the center of the first community, patent text similarity and hierarchical agglomerative
clustering analysis was undertaken. These methods are of great use to discover the groups
13 This number includes patents that cited 3GPP standards together with other formal standards.
Page 22
21
of firms with similar technological capabilities. Since firms with similar patents tend to
interact in a strategically competitive manner, the finding of those groups helps us to
comprehend the dynamics of M2M/IoT innovation. We used the text of granted patents that
center on the M2M/IoT trajectory and patents that cited, and were cited by, the M2M/IoT
trajectory patents. Duplicate patents were removed. The text of 3GPP standards was also
included. To find a change in the M2M/IoT technology clusters, patents were split into two
sets: one that represented the earlier M2M/IoT trajectory from 2008 to 2010 (Ericsson
(8407769), Apple (8737989), InterDigital (8718688)), and the other that represented the latter
from 2010 to 2012 (InterDigital (8718688), Qualcomm (9210527), Alcatel Lucent (9071925),
HTC (9167470), LG (9137624), Huawei (9107226), Sprint (8638724)).
For the former set, using Ward’s hierarchical agglomerative clustering method, we
extracted four clusters around the earlier M2M/IoT trajectory, as shown in Table 6. Ward’s
hierarchical agglomerative clustering method focuses on node similarity in lieu of the
sharing of the same edges. The first cluster contains 55 patents, including two trajectory
patents (Ericsson (8407769), InterDigital (8718688)). In the first cluster, as being the largest
group, several different firms’ technologies are positioned and their share of patents is
relatively evenly distributed. Samsung, Qualcomm, Nokia, Jasper and Huawei, inter alia, are
main industry players, selected by their share of patents within the cluster. This group of
firms is in potentially cooperative and competitive relations with respect to M2M/IoT
technological innovation. Other three clusters exhibit skewed distributions. Lemko, Ericsson
and LG are dominant players in the second, third and fourth clusters, respectively. Lemko is
a US-headquartered firm that provides quick deployments of LTE cellular systems with low
cost, powered by virtualized and distributed EPC (Evolved Packet Core) and IMS (IP
Multimedia Subsystem) solutions (Lemko, 2013). Interestingly in the third cluster, Ericsson
and InterDigital have a large portion of patents that are compatible with 3GPP standards.
Using the mean of each cluster’s patent similarity, we applied a hierarchical
agglomerative clustering method and pairwise comparisons, as shown in Figure 5. At a
higher level, the first and fourth clusters, which include the M2M/IoT trajectory patents, are
grouped as the same cluster (correlation coefficient = 0.81, p-value < 0.01). The third cluster
(3GPP standards) is also similar to the first cluster (main cluster) (correlation coefficient =
0.34, p-value < 0.01). This shows that technologies centering around the main M2M/IoT
trajectory are compatible with 3GPP standards. The fourth cluster is statistically significant
similar with the first cluster and simultaneously not similar with the third cluster. This
finding shows that LG, a dominant player in the fourth cluster, has attempted to create a
new technological path that is distinctive with the main M2M/IoT trajectory.
Table 6. Firms which own patents in each cluster around the earlier M2M/IoT trajectory
Cluster 1
(55 patents)
Cluster 2
(30 patents)
Cluster 3
(31 patents and 6
standards)
Cluster 4
(32 patents)
Trajectory:
Ericsson (8407769),
InterDigital (8718688)
Standards:
3GPP TS 23.003, TS
23.008, TS 23.060, TS
43.020, TR 22.868, TR
33.812
Trajectory:
Apple (8737989)
Samsung (12.7%) Lemko (36.7%) Ericsson (37.8%) LG (28.1%)
Page 23
22
Qualcomm (7.3%) Alcatel-Lucent (6.7%) InterDigital (10.8%) Apple (9.4%)
Nokia (7.3%) Motorola (6.7%) M2M and IoT (5.4%) Vodafone (6.3%)
Jasper (7.3%) AT&T (6.7%) Apple (2.7%) Intel (6.3%)
Huawei (5.5%) Ericsson (6.7%) Alcatel-Lucent (2.7%) Cellco (6.3%)
Note: Top five firms were selected based on their share of patents in each cluster. Firms’ shares are in
parentheses.
Figure 5. Patent clusters around the earlier M2M/IoT trajectory (Left) and correlations between the clusters
(Right). Note: In the heatmap, each row indicates each cluster, while each column indicates each patent. A color
closer to yellow represents similarity, whereas a color closer to red represents dissimilarity.
As for the latter M2M/IoT trajectory, we detected four clusters, as shown in Table 7. The first
cluster is composed of 46 patents, including five trajectory patents (InterDigital (8718688),
LG (9137624), Sprint (8638724), HTC (9167470), Alcatel Lucent (9071925)). This cluster can be
regarded the main strategic group in the main cluster of the latter trajectory. Compared to
the strategic group positions in the main cluster of the earlier trajectory, the positions of
Nokia, LG and Alcatel-Lucent in the main strategic group become more central with the
passage of time. InterDigital and Huawei serve as the main players in the second cluster,
while Qualcomm holds an influential position in the third cluster. Akin to the earlier
trajectory, Ericsson owns a distinctly large portion of patents in the fourth cluster, which are
similar with the text of 3GPP standards.
Between-cluster similarities are presented in Figure 6. The similarity of the first
cluster with three other clusters is all relatively high. This confirms that the first cluster
serves as the main cluster, connecting three other clusters. It is notable that the fourth cluster
(3GPP standards) is compatible with the main cluster (correlation coefficient = 0.33, p-value
< 0.01), whereas its correlations with other two clusters (second and third) are not
statistically significant at the 5% level. It implies that Qualcomm, InterDigital and Huawei
have made efforts to generate new technological innovations that are differentiable from the
3GPP standards-based technology.
Page 24
23
Table 7. Firms which own patents in each cluster around the latter M2M/IoT trajectory
Cluster 1
(46 patents)
Cluster 2
(17 patents and
1 standard)
Cluster 3
(39 patents)
Cluster 4
(10 patents and
5 standards)
Trajectory:
InterDigital (8718688),
LG (9137624),
Sprint (8638724),
HTC (9167470),
Alcatel Lucent
(9071925)
Trajectory:
Huawei (9107226)
Standards:
3GPP TS 23.003
Trajectory:
Qualcomm (9210527)
Standards:
3GPP TS 23.107, TS
22.367, TS 22.368, TS
23.060, TR 23.888
Nokia (10.9%) InterDigital (16.7%) Qualcomm (10.3%) Ericsson (53.3%)
LG (8.7%) Huawei (11.1%) Motorola (7.7%) Samsung (6.7%)
Alcatel-Lucent (8.7%) M2M and IoT (11.1%) Huawei (5.1%) Alcatel-Lucent (6.7%)
Qualcomm (6.5%) Fujitsu (5.6%) Samsung (5.1%)
Samsung (6.5%) Ericsson (5.6%) NTT Docomo (5.1%)
Note: Top five firms were selected based on their share of patents in each cluster. Firms’ shares are in
parentheses.
Figure 6. Patent clusters around the latter M2M/IoT trajectory (Left) and correlations between the clusters (Right)
4.4. Hypothesis test results
To empirically examine the effects of formal standards on the basicness of patents, statistical
analysis was conducted to test the aforementioned hypotheses (H1, H2, H3). Table 8 shows
the test results for H1. In models 1, 2, 3, 4 (Tobit regression), a reference to standards (the
main IV) is positively associated with adjusted forward citations (DV) with statistical
significance at the 0.01 level. This test result is robust to the violation of the homoscedasticity
and normality assumptions, as shown in model 5 (CLAD regression). The robustness of the
result is reconfirmed by the negative binomial regression test (model 6). Hence, the first
hypothesis (H1) is supported.
Page 25
24
Table 8. Empirical test results for the first hypothesis (H1)
Model 1 (AC) Model 2 (AC) Model 3 (AC) Model 4 (AC) Model 5 (AC) Model 6 (FC)
Tobit Tobit Tobit Tobit CLAD Negative
Binomial
Standards 0.531** 0.578** 0.574*** 0.623*** 0.357** 0.367**
Backward
citations
0.022*** 0.018*** 0.017*** 0.017*** 0.017*** 0.006***
Catch-up -0.258 -0.358 -0.575** -0.482* -0.108 -0.420**
Year dummy No No No Yes No Yes
LnFirmSize -0.099 -0.135 0.010 -0.067
LnIP 0.154 0.167 0.050 0.130*
LnR&D -0.240* -0.212 -0.091 -0.184*
Constant 0.258 -0.106 5.022*** 5.531*** 1.496 6.015***
Observations 421 343 343 343 343 343
Log-
likelihood
-664.561 -534.904 -527.890 -525.313 -676.421
Pseudo R2 0.113 0.043 0.055 0.060 0.082 0.065
*** p < 0.01, ** p < 0.05, * p < 0.1. AC: Adjusted Forward Citations. FC: Forward Citations. Standard errors in parentheses.
Note: Patents that were assigned by small and medium-sized enterprises whose firm-specific information is not publically
available were excluded from the analysis in models 2, 3, 4, 5, 6.
Table 9 shows the test results for H2. In model 1 (Tobit), a reference to standards (IV) is
negatively associated with the generality of patents with statistical significance at the 0.05
level. A robustness check was also conducted through the CLAD estimator. As shown in
models 4 and 5, it validates the test results, supporting the second hypothesis (H2). It is
noticeable that the statistical significance of the coefficients on standards drops when we
add originality and originality-squared as IVs, as shown in models 2 and 3. This leads us to
conjecture the existence of the moderating effect of standards on the association between
originality and generality. The test results in models 3 and 5 show that originality and
generality are related with statistical significance in a curvilinear manner.
Table 9. Empirical test results for the second hypothesis (H2)
Model 1 (GE) Model 2 (GE) Model 3 (GE) Model 4 (GE) Model 5 (GE)
Tobit Tobit Tobit CLAD CLAD
Standards -0.074** -0.061* -0.045 -0.085** -0.068**
Originality 0.225** -0.678** 0.301** -0.858*
Originality2 0.878*** 1.094**
Forward citations 0.011*** 0.009*** 0.009*** 0.006* 0.007***
Page 26
25
Catch-up -0.168*** -0.149*** -0.147*** -0.126*** -0.138***
Constant 0.461*** 0.308*** 0.484*** 0.330*** 0.575***
Observations 253 253 253 253 253
Log-likelihood -65.389 -62.159 -58.293
Pseudo R2 0.277 0.312 0.355 0.116 0.145
*** p < 0.01, ** p < 0.05, * p < 0.1. GE: Generality. Standard errors in parentheses.
For hypothesis H3, we examined the moderating effect of standards on the relationship
between binding force and generality. We added the interaction term between standards
and binding force to the test model for generality. The interaction term between standards
and originality was also added, since the previous test results (Table 9) indicated the
existence of the moderating effect of standards on the association between originality and
generality. Forward citations were added as a control variable in the Tobit model in order to
tease out the effect of binding force from the build-in coverage effect of a large number of
forward citations.
Table 10 shows the moderation test results for H3. The coefficients on the interaction
terms between standards and binding force and between standards and originality are
statistically significant, supporting H3. A multicollinearity test was also conducted by
checking variance influence factors (VIF) (Mason & Perreault, 1991). The test showed that all
VIF values were less than 5, indicating that multicollinearity was not likely to distort the test
results of this study. Table 10. Empirical test results for the third hypothesis (H3)
Model 1 (GE) Model 2 (GE) Model 3 (GE) Model 4 (GE)
Tobit Tobit Tobit Tobit
Standards -0.043 -0.470*** -0.031 -0.486***
Originality -0.603* -1.758*** -0.484 -1.611***
Originality2 0.810*** 1.907*** 0.668** 1.700***
Binding force 0.004 0.006 -0.026* -0.020
Forward citations 0.008** 0.007** 0.021*** 0.018***
Standards × Originality 2.131*** 2.010***
Standards × Originality2 -2.078*** -1.866***
Standards × Binding force 0.069** 0.064**
Standards × Forward citations -0.017** -0.014*
Catch-up -0.148*** -0.142*** -0.152*** -0.143***
Constant 0.466*** 0.702*** 0.428*** 0.683***
Observations 253 253 253 253
Log-likelihood -58.488 -52.354 -54.750 -49.668
Page 27
26
*** p < 0.01, ** p < 0.05, * p < 0.1. GE: Generality. Standard errors in parentheses.
For a categorical moderator, the coefficients on the interaction terms are the differences of
the regression coefficients between two subgroups: one without a reference to standards
(STD=0) and the other with a reference to standards (STD=1). Subgroup analysis was
conducted to examine the moderating effect of standards (Sharma et al., 1981). The graphical
representation of the moderating effect of standards is presented in Figure 7.
Figure 7. Graphical representation of the moderating effect of standards
Figure 8 shows the results of the pairwise comparisons of the relationships among
originality, binding force and generality. In the first subgroup (STD=0), it is shown that there
is a curvilinear relationship between originality and generality. There is also a noticeable
difference between the two subgroups in the correlation coefficient of the pair of originality
and binding force (0.11 for STD=0, -0.11 for STD=1). As compared to the second subgroup
(STD=1), the binding force in the first subgroup (STD=0) is pulled by an outlier, as shown in
Figure 9. This outlier is Seven Networks’s patent 9060032 (selective data compression by a
distributed traffic management system to reduce mobile data traffic and signaling traffic),
which was regarded as one of the strategic patents of M2M/IoT in Table 4.
Figure 8. Scatterplots, histograms, correlations for pairwise comparison matrices of originality, binding force and
generality. STD=0 (Left) and STD =1 (Right).
Page 28
27
Figure 9. Scatterplot3D of binding force, originality, generality
One more finding is that the patents filed by catch-up firms are less cited by subsequent
patents and more specific in the technological field, as compared to firms in other countries
(mainly Europe and US), as shown in Table 8 and 9. This is in line with the findings of
previous research (e.g. Kang et al., 2014; Park & Lee, 2006) that catch-up firms tend to lack
technological resources and capabilities to generate highly valuable technologies and hinge
on foreign knowledge, particularly from European and US firms, and thereby selectively
focus on a specific technological field. Since the emergence of a new technological paradigm
opens a window of opportunity to catch up (Perez & Soete, 1988), latecomer firms, especially
from Korea and China, have been vigorously utilizing formal standards of new technologies
(ICT in particular) in the catch-up context (K. Lee & Lim, 2001; Yu, 2011). This is confirmed
by the independent t-test for a difference in means with respect to standards between catch-
up firms and the others (mostly European and US firms). Approximately 61% of the catch-
up firm group referenced to formal standards, while 44% of the other group cited standards.
This difference is statistically significant (t = 3.298, p < 0.01; equal variances not assumed).
Figure 10 shows an increase in catch-up firms’ M2M/IoT technologies over time. Catch-up
firms tend to rely on formal standards, and thereby their technologies are more concentrated
on the center of the patent network.
Page 29
28
Figure 10. Changes in catch-up firms’ M2M/IoT technologies (colored green) in the network of patents with high
betweenness centrality (top 5%). Application year 1999-2009 (Left) and 2010-2015 (Right). Node size represents
generality.
5. Discussion
5.1. Standards as a driving force of technological convergence
The analysis of the M2M/IoT patent citation network in this paper shows that formal
standards affect the clustering of patents with high betweenness centrality and the shaping
of a main technological trajectory. This serves as clear evidence demonstrating that formal
standards guide the direction of technological change. It is possible because standards
constitute the essence of a technological paradigm by which the meanings of technical
artifacts and their relational properties are stabilized. This finding is of particular
significance in the studies of innovation and standardization since there has been no prior
research that identifies a main technological trajectory that has been shaped by formal
standards.
According to the results of hypotheses tests, as shown in Tables 8 and 9,
technologies based on standards are likely to be more valuable and specific. The
confirmation of the first hypothesis is of significance since, to the best of our knowledge, it is
the first paper that statistically demonstrates the relationship between a reference to
standards and the importance of patents. This accentuates the role of standards as a base for
future technological development, particularly in the field of network technologies.
The validation of the second hypothesis that a reference to standards is negatively
associated with the generality of patents needs to be carefully read with Cassiman et al.'s
(2008) finding that a citation to scientific publications is positively related with the scope of
forward citations. Patent citations links to scientific literature have been regarded as an
indicator of a highly mediated interaction between science and technology (Meyer, 2000). As
a map for technological landscapes, scientific knowledge guides inventors in the direction of
new combinatory exploration rather than local search (exploitation) (Fleming & Sorenson,
2004). This explains the positive effect of scientific linkages on the generality of technology
(Cassiman et al., 2008). Standards are, by definition, different from scientific knowledge.
Page 30
29
Standards attempt to specify an optimized version of technological variants, whereas science
focuses on the generalization of research. That is to say, standards’ function of variety
reduction is key to understand the different effects of science and standards on innovation.
The effects of standards on technological variants are complex since there is the
multifaceted and long-range interplay between standardization and innovation (Foster &
Heeks, 2013; Zoo, Vries, & Lee, 2017). Standards as dominant designs drive technological
progress into the era of incremental innovations (Anderson & Tushman, 1990). This stability
continues until the punctuation by a new technological discontinuity that accelerates the rate
of technological variation. In ICT standardization, modular designs have been employed to
address this inherent tension between stability and flexibility (Hanseth et al., 1996). By
setting architectural structures and common interfaces, 3GPP standards enable different
subsystems to operate in a compatible manner and offer a foundation for innovations in the
long term. This foundational knowledge tends to be frequently cited by subsequent
inventions, and gives rise to convergence in complements over time (Greenstein & Khanna,
1997).
The moderating effect of standards on the relationships among binding force,
originality and generality, as shown in Table 10 and Figure 7, needs to be factored in the
understanding of the impact of standards on technological variations within the context of
technological convergence. Discussions on the convergence of computing and
telecommunications systems have a long history—for instance, Science Magazine covered
David Farber and Paul Baran’s article regarding this convergence issue in 1977 (Farber &
Baran, 1977).14 Yoffie (1997) identified three drivers that precipitated these trends:
technological factors (semiconductor, software and digital communications), government
deregulation and managerial creativity. Technological convergence resulted in industry
convergence from vertical integration towards a horizontal structure, which required a high
degree of coordination and common interfaces. Despite the significant role of standards in
industry convergence and value creation, acknowledged by many researchers (e.g. Jacobides,
Knudsen, & Augier, 2006; McGahan, Vadasz, & Yoffie, 1997), previous research has not yet
empirically demonstrated the effect of standards as a driving force of technological
convergence. The discovery of the moderating effect of standards in this paper fills this
research gap.
Prior empirical studies on technological convergence viewed binding forces and
technological diversity as crucial factors to investigate convergence across the technological
fields (Y. Cho & Kim, 2014; Han & Sohn, 2016). In these studies, high levels of binding forces
and technological diversity were deemed as results of technological convergence. Our
research added two more factors (i.e. standards and originality) to examine the relationship
between binding forces and generality at a technological patent level. According to our
empirical test results, the generality of patents without reference to standards is highly
dependent on their originality. After passing a certain threshold, patents citing prior art
from various fields tend to be cited by subsequent patents from a wide range of fields. Yet
binding forces, instead of originality, hold a more significant role in technological diversity
for the patents that build upon standards. This means that standards with specific
14 Dr. Farber and Mr. Baran were regarded as pioneers who made significant contributions to the development of
computer networks (e.g. Packet switched network design and Token Ring).
Page 31
30
architectural knowledge and compatibility requirements serve first as a force of
concentration that underpins the construction of a large, densely connected community of
M2M/IoT technologies. In this community, nodes with high central positions work as a
source of attraction that brings relatively heterogeneous technologies.
5.2. Importance of 3GPP standards for M2M/IoT technological evolution
One of the crucial findings in this paper is that 3GPP standards, among formal standards,
are most frequently referenced and, in turn, heavily affect the M2M/IoT technological
trajectory. Due to high switching costs, investments on a particular standard involve
substantial risks, which critically affect the performance of firms in highly competitive
technology markets (Cusumano, Mylonadis, & Rosenbloom, 1992; Shapiro & Varian, 1999b).
Multiple variants of M2M/IoT standards are currently vying to be selected in the standards-
setting committees and become dominant in the market—iner alia, 3GPP, ETSI, IETF, IEEE,
AllSeen, OIC, Thread, and IIC. Despite the fact that those committees and consortia attempt
to differentiate themselves from others in the field of M2M/IoT standardization, some of
their standardization efforts overlap and, in turn, come into inexorable competition. In this
context, the finding that, among formal standards, 3GPP standards are the one that most
strongly shapes the M2M/IoT technological trajectory is particularly relevant for senior
managers in the formation of technology strategy.
Standardization efforts within 3GPP have focused upon the optimization of
core/access networks for M2M traffic15 and development of high-level frameworks and
service capabilities. Accordingly, 3GPP TR 22.868, TR 33.812, TS 22.368 and TR 23.888 were
released in line with those efforts. TR 22.868 (study on facilitating machine to machine
communication in 3GPP systems) was first drafted in 2006, identifying potential
requirements to facilitate the optimization of radio and networks resources in the followings
areas: inter alia, handling large numbers of terminals and subscription data, charging,
security and addressing. TR 33.812 (feasibility study on security aspects), which became
available since 2008, addresses issues regarding remote provisioning and change of
subscription for M2M equipment. TS 22.368, first available in 2009, specifies service
requirements for M2M communications, which serve as a foundation for architecture and
protocol specification in other 3GPP groups. In this standard, a distinction between
common-service requirements and M2M features that only apply to a specific subscription is
made, predicted upon the idea that telecom operators can differentiate their services on a
per-subscription basis (Norp & Landais, 2012). TR 23.888, released in 2009, studies and
evaluates the architectural aspects of system improvements for M2M requirements, specified
in TS 22.368.
The main flow of knowledge, embodied in the top path of the M2M/IoT patent
citation network, reflects a stream of the aforementioned 3GPP standardization efforts.
US8407769 (Ericsson, 2008) contains methods that facilitate the automatic linking of a newly
activated M2M devices to an appropriate server for downloading subscription credentials.
US8737989 (Apple, 2008) includes methods to enable a wireless network to identify
15 There are differences between M2M traffic and H2H (human-to-human) traffic: 1) synchronized; 2)
unpredictable; 3) bursty; and 4) uncontrollable. For details, see Benrachi-Maassam (2012).
Page 32
31
subscriptions and then offer differentiated services to a M2M client based on this
identification. US8718688 (InterDigital, 2010) provides methods for solving limited
addressing space in M2M environments where communications between a controller and a
group of transmit/receive units, having a same international mobile subscriber identity
(IMSI), take place. It is noticeable that the M2M/IoT technological trajectory in line with
3GPP standardization efforts has been taking shape in response to well-recognized needs in
the market (i.e. a smart, connected product as a service (Porter & Heppelmann, 2014)). The
product-as-a-service model can be maintained through the provision of differentiated
services on the basis of their subscription data. This flow of discussion leads to another
finding that the direction of a path-dependent technological trajectory is influenced by
market demand-embodied formal standards.
5.3. Two different types of technology clusters (standards-based and platforms-based
clusters)
The discovery of technology clusters in the M2M/IoT patent network shows the groups of
firms with similar M2M/IoT technological capabilities. This finding is of significance
especially during the era of M2M/IoT-driven industry convergence where the boundaries of
traditional industrial sectors become blurred and, in turn, convergences in complements
(use in concert) and in substitutes (interchangeability) become growingly important
(Greenstein & Khanna, 1997). Bound to a common fate in the face of uncertainty, firms with
similar technological capabilities in the strategic groups are likely to create different
“strategic blocks” and compete with one another (complementary blocks composed of firms
from different strategic groups and pooling blocks composed of firms from the same
strategic group) (Nohria & Garcia-Pont, 1991). This group-based competition is critical in
value creation and battles over technological standards (Gomes-Casseres, 1994; McGahan et
al., 1997).
Those firms with similar M2M/IoT patents are likely to interact in a cooperative and
competitive (coopetitive) manner in their value networks. Cooperation takes place among
complementors in order to create value (a larger pie), whereas competition occurs in a way
to appropriate value (a larger slice) (Brandenburger & Nalebuff, 1996). In the face of
technological challenges and opportunities (e.g. the setting of technological standards), giant
firms actively engage in coopetition, which results in subsequent coopetition among other
firms and the advance of innovation (Gnyawali & Park, 2011). In the formation of strategic
linkages, firms’ resources (e.g. accumulated technological assets and radical technological
breakthroughs) and their network positions (crowding and prestige) exert a gravitational
pull (Ahuja, 2000; Stuart, 1998). Firms in the central position (in terms of betweenness) of a
concentrated interfirm network (in terms of density) are more likely to successfully develop
explorative innovations (Gilsing, Nooteboom, Vanhaverbeke, Duysters, & van den Oord,
2008).
In our analysis of patent text similarity (Tables 6 and 7), Ericsson, InterDigital,
Samsung, Qualcomm, Nokia, Jasper, Huawei, LG, and Alcatel-Lucent are main firms in the
same strategic group composed of similar technological capabilities that are compatible with
3GPP standards. Those firms are likely to intensely interact in a cooperative and competitive
manner in order to create and appropriate value in the face of M2M/IoT standards-driven
convergence. Standards-based alliances have been strategic responses to the growing
Page 33
32
complexity of ICT-driven technological systems (Rice & Galvin, 2006). The texts of Ericsson’s
patents are most similar to those of 3GPP standards. Interestingly, InterDigital, Huawei, and
Qualcomm created different technology clusters that were less similar to 3GPP standards
along the M2M/IoT trajectory over time. It implies that those firms attempted to diversify
3GPP standards-based M2M/IoT technologies, generating new paths of the M2M/IoT
trajectory.
As shown in the M2M/IoT technological communities (Table 3), Google and
Amazon carved out their self-reliant technological trajectory through acquisition of
strategically important patents, instead of relying on formal standards. This market-oriented
approach can be regarded as part of “platform envelopment”, which refers to an entry into
adjacent markets by bundling their own platform’s functionality with that of other products
and services, leveraging common components and overlapping user bases (Eisenmann,
Parker, & Alstyne, 2011). Envelopment opportunities for platform leaders arise with a large-
scale industry convergence which allows them to compete in a multi-layered technological
space with different platforms (Eisenmann, Parker, & Alstyne, 2006). Platform leaders often
use envelopment as a “tipping” strategy to build market momentum and win a platform
competition (Gawer & Cusumano, 2008).
Google has made inroads into many different markets by connecting new
technological features to its search platform, innovatively linked with an ad-based business
model (Kenney & Pon, 2011). In January 2014, Google acquired Nest’s learning thermostat
and absorbed technological knowledge regarding a control unit in the network of home
appliances (The Economist, 2014). By combining new technological features building upon
this stock of knowledge with existing platforms and services, Google attempted to create its
own IoT ecosystem where cross-platform interoperability was ensured to reduce the costs of
multi-homing (i.e. affiliating with multiple platforms (Eisenmann et al., 2006)). In turn, users
were able to enjoy a plethora of platform-agnostic services, whereas still locked in Google’s
search engine (i.e. the core of the Google ecosystem, funneling user data into its database to
produce, accumulate, and re-appropriate value (Pasquinelli, 2009)). The accumulated
experience of a user interface affects the way switching costs results in user resistance to
change, and thereby works as a critical factor in multi-platform competition (D. Kim & Lee,
2016). In this context, the creation of a user-friendly interface for control unit is crucial for
Google’s tipping strategy, as shown in its patent 9223323.
Likewise, Amazon has been trying to offer various web services on top of its own
platform where all the transaction data of users is stored in its database to capture value. Its
efforts to obtain Benjamin’s M2M/IoT patents can be interpreted within the context of an
envelopment attempt to connect household gadgets to Amazon’s existing functionality and
services, building on its e-commerce platform and cloud infrastructure. A platform owner
emphasizes the role of a gatekeeper which allows the firm to exercise control over its
platform and, in turn, focuses on the development of gatekeeper functions, such as
profile/identity management, service provisioning/service brokerage, and charging and
billing (Ballon, 2009). The control of customer identity information that can be charged for
viewing contents is critical for Amazon’s gatekeeper strategy, as shown in its patent 8666308.
There is a noticeable pattern in the different technology clusters of the M2M/IoT
patent network. Previous research identified changes in industry structure in the face of
technological convergence as a horizontal structure (chips, computer, operating system,
application software and distribution in the computer industry (Grove, 1996) and terminals,
Page 34
33
manipulation, transmission, packaging and content in the converging industry of
telecommunications, computing and entertainment (Collis, Bane, & Bradley, 1997)).
According to our analysis, firms previously involved in terminal and transmission sectors
tended to develop similar technological capabilities, which were compatible with formal
standards (e.g. Samsung, Apple, LG, Huawei, Ericsson, and Nokia). Firms in manipulation
and packaging sectors honed in on self-reliant technologies that were essential for their
platform envelopment and gatekeeping roles (e.g. Google and Amazon).
5.4. Standards for a path-creating catch-up and Huawei’s rise
Our findings confirm that the M2M/IoT technologies of catch-up firms are more
concentrated on a standards-based cluster, as shown in Figure 10. It appears that catch-up
firms play a growingly important role in the creation of a new path along the M2M/IoT
trajectory. It is particularly noticeable that Huawei initiated a path-creating catch-up. The
finding of patent text similarity (Table 7) shows that Qualcomm, InterDigital and Huawei
actively engaged in technological diversification, distinguishable from other 3GPP
standards-based technologies. Previously, the knowledge positions of Qualcomm and
InterDigital only were considered strong in standards-based markets (Bekkers & Martinelli,
2012). In our analysis, Huawei’s role in the shaping of the M2M/IoT trajectory is growingly
influential. For instance, Huawei’s patent 9107226 (congestion handing in a communication
system) referenced to the IEEE 802.16 standard (WiMax). This patent is cited by HTC’s
patent 9167470 (signaling congestion handling), which referenced to 3GPP standards. This
path of connecting two different standards-based technologies serves as an important part of
the M2M/IoT trajectory. This finding stresses the significance of standards as a post catch-up
strategy.
In the analysis of our dataset, over 70% of Huawei’s M2M/IoT patents referenced to
3GPP standards, especially TR 22.868, TR 33.812, TS 22.368 and TR 23.888. By contrast, ZTE,
another Chinese catch-up firm, has relied relatively less on formal standards. Only 35% of
ZTE’s M2M/IoT patents cited to 3GPP standards-related documents, some of which were
just 3GPP working group papers. This shows that Huawei has been making innovative
efforts, more squarely related to the 3GPP-shaped M2M/IoT technological trajectory, in
comparison to ZTE. This implies that Huawei already recognized the strategic importance of
3GPP standards with respect to an emerging M2M/IoT technology and, in turn, vigorously
built their technological capabilities in line with these standards.
6. Conclusion
With a growing strategic importance of understanding the M2M/IoT trajectory and its
relationship with formal standards, this research offers several findings and
theoretical/practical implications. First, our empirical analysis shows that standards serve as
a driver of technological convergence. Second, we found that 3GPP standards regarding the
overall M2M/IoT system architecture (TR 22.868, TR 33.812, TS 22.368, and TR 23.888, in
particular) assume a leading role in the shaping of the M2M/IoT trajectory. Third, we
identified strategic groups and strategic patents centering around the M2M/IoT trajectory.
This identification is crucial to understand the dynamics of strategic competition in the
Page 35
34
M2M/IoT standards-driven converging world of Internet and things. Forth, standards serve
as a critical factor in the process of creating a new path for catch-up firms (e.g. Huawei).
There findings are expected to provide contributions to innovation and standards
studies by empirically investigating the relationship between technological trajectories and
standards. Yet these findings should be interpreted with care. There is a possibility that a
reference to 3GPP standards is the most decisive factor for US patent examiners to categorize
M2M/IoT patents into CPC code H04W/005, which may result in a bias in the finding of
3GPP standards’ significant impacts on the M2M/IoT trajectory. Despite the fact that we
cannot completely rule out this possibility, the existence of the non-H04W/005 patents citing
those 3GPP standards and the H04W/005 patents not citing the 3GPP standards indicates
that the probability of this factor causing a bias in the finding is not great. For future
research, it would be relevant to further investigate the dynamics of strategic competition in
the standardization process of M2M/IoT-related 3GPP standards.
There are some limits on this research. First, we relied on several different
algorithms to detect technology clusters in the M2M/IoT networks, including non-
deterministic, approximation methods (e.g. Blondel’s et al.’s modularity optimization). Non-
deterministic approach hinges on initial conditions and/or parameters of the algorithm, and
thus may not deliver an exactly same solution to the same problem. In order to overcome
this weakness, we attempted to use different methods to check the robustness of our
findings. For instance, the CW algorithm was used to support the technology clusters in the
M2M/IoT network, identified by Blondel’s et al.’s algorithm. Node betweeneess centrality
was also employed to confirm the central positions of trajectory patents in the M2M/IoT
network. Second, there are insufficient explanations on the characteristics of each cluster in
the M2M/IoT network, despite some explanations in the paper—for example, these
technology clusters emerged based on the sharing of similar information (e.g. 3GPP
standards) or the group of firms with strategic interest (e.g. Google and Nest Labs). This
leaves open room for future research on the emergence of different technology clusters in
the M2M/IoT network.
Acknowledgement
This work was supported by the National Research Foundation of Korea Grant funded by
the Korean Government (NRF-2014S1A3A2043505) and ICONS (Institute of Convergence
Science), Yonsei University.
References
Ahuja, G. (2000). The duality of collaboration: Inducements and opportunities in the
formation of interfirm linkages. Strategic Management Journal, 21(3), 317.
http://doi.org/10.2307/3094190
Anderson, P., & Tushman, M. L. (1990). Technological discontinuities and dominant designs:
A cyclical model of technological change. Administrative Science Quarterly, 35(4), 604–
633. http://doi.org/10.2307/2393511
Arthur, W. B. (1994). Increasing Returns and Path Dependence in the Economy. University of
Michigan Press.
Page 36
35
Athreye, S., & Keeble, D. (2000). Technological convergence, globalization and ownership in
the UK computer industry. Technovation, 20(5), 227–245. http://doi.org/10.1016/S0166-
4972(99)00135-2
Atzori, L., Iera, A., & Morabito, G. (2010). The Internet of Things: A survey. Computer
Networks, 54(15), 2787–2805. http://doi.org/10.1016/j.comnet.2010.05.010
Ballon, P. (2009). The platformisation of the European mobile industry. Communications &
Strategies, (75), 15–34.
Barberá-Tomás, D., Jiménez-Sáez, F., & Castelló-Molina, I. (2011). Mapping the importance
of the real world: The validity of connectivity analysis of patent citations networks.
Research Policy, 40(3), 473–486. http://doi.org/10.1016/j.respol.2010.11.002
Bekkers, R., Bongard, R., & Nuvolari, A. (2011). An empirical study on the determinants of
essential patent claims in compatibility standards. Research Policy, 40(7), 1001–1015.
http://doi.org/10.1016/j.respol.2011.05.004
Bekkers, R., & Martinelli, A. (2012). Knowledge positions in high-tech markets: Trajectories,
standards, strategies and true innovators. Technological Forecasting and Social Change,
79(7), 1192–1216. http://doi.org/10.1016/j.techfore.2012.01.009
Bekkers, R., Verspagen, B., & Smits, J. (2002). Intellectual property rights and
standardization: the case of GSM. Telecommunications Policy, 26(3), 171–188.
Benrachi-Maassam, S. (2012). Lessons learned from early M2M developments. In D.
Boswarthick, O. Elloumi, & O. Hersent (Eds.), M2M Communications: A Systems
Approach (pp. 57–71). John Wiley & Sons.
Bergmann, I., Butzke, D., Walter, L., Fuerste, J. P., Moehrle, M. G., & Erdmann, V. A. (2008).
Evaluating the risk of patent infringement by means of semantic patent analysis: The
case of DNA chips. R&D Management, 38(5), 550–562. http://doi.org/10.1111/j.1467-
9310.2008.00533.x
Biemann, C. (2006). Chinese whispers: an efficient graph clustering algorithm and its
application to natural language processing problems. In Workshop on Text Graphs at
HLT-NAACL (pp. 73–80).
Bijker, W. E. (1995). Of Bicycles, Bakelites, and Bulbs: Toward a Theory of Sociotechnical Change.
MIT Press.
Blind, K. (2004). The Economics of Standards: Theory, Evidence, Policy. Edward Elgar Publishing.
Blind, K., & Thumm, N. (2004). Interrelation between patenting and standardisation
strategies: Empirical evidence and policy implications. Research Policy, 33(10), 1583–1598.
http://doi.org/10.1016/j.respol.2004.08.007
Blondel, V. D., Guillaume, J.-L., Lambiotte, R., & Lefebvre, E. (2008). Fast unfolding of
communities in large networks. Journal of Statistical Mechanics: Theory and Experiment, 10.
http://doi.org/10.1088/1742-5468/2008/10/P10008
Brandenburger, A. M., & Nalebuff, B. J. (1996). Co-opetition: A Revolution Mindset that
Combines Competition and Cooperation. Currency Doubleday.
Burt, R. S. (1992). Structural Holes: The Social Structure of Competition. Harvard University
Press.
Cameron, A. C., & Trivedi, P. K. (1998). Regression Analysis of Count Data. Cambridge
University Press. Retrieved from papers3://publication/uuid/3EDE35CA-32B8-49A6-
92AE-E85361BEEBB0
Page 37
36
Carpenter, M. P., Narin, F., & Woolf, P. (1981). Citation rates to technologically important
patents. World Patent Information, 3(4), 160–163. http://doi.org/10.1016/0172-
2190(81)90098-3
Cassiman, B., Veugelers, R., & Zuniga, P. (2008). In search of performance effects of
(in)direct industry science links. Industrial and Corporate Change, 17(4), 611–646.
http://doi.org/10.1093/icc/dtn023
Chen, M., Wan, J., & Li, F. (2012). Machine-to-machine communications: Architectures,
standards and applications. KSII Transactions on Internet and Information Systems, 6(2),
480–497. http://doi.org/10.3837/tiis.2012.02.002
Cho, T. S., & Shih, H. Y. (2011). Patent citation network analysis of core and emerging
technologies in Taiwan: 1997-2008. Scientometrics, 89(3), 795–811.
http://doi.org/10.1007/s11192-011-0457-z
Cho, Y., & Kim, M. (2014). Entropy and gravity concepts as new methodological indexes to
investigate technological convergence: Patent network-based approach. PLoS ONE, 9(6).
http://doi.org/10.1371/journal.pone.0098009
Clauset, A., Newman, M. E. J., & Moore, C. (2004). Finding community structure in very
large networks. Physical Review E, 70(6).
Collis, D. J., Bane, P. W., & Bradley, S. P. (1997). Winners and losers: Industry structure in
the converging world of telecommunications, computing, and entertainment. In D. B.
Yoffie (Ed.), Competing in the Age of Digital Convergence (pp. 159–200). Harvard Business
School Press.
Cusumano, M. A., Mylonadis, Y., & Rosenbloom, R. S. (1992). Strategic maneuvering and
mass-market dynamics - the triumph of VHS over Beta. Business History Review, 66(1),
51–94.
David, P. A., & Bunn, J. A. (1988). The economics of gateway technologies and network
evolution: Lessons from electricity supply history. Information Economics and Policy, 3,
165–202.
David, P. A., & Greenstein, S. (1990). The economics of compatibility standards: An
introduction to recent research. Economics of Innovation and New Technology, 1, 3–41.
David, P. A., & Steinmueller, W. E. (1994). Economics of compatibility standards and
competition in telecommunication networks. Information Economics and Policy, 6(94),
217–241. http://doi.org/10.1016/0167-6245(94)90003-5
de Vries, H., Verheul, H., & Willemse, H. (2003). Stakeholder identification in IT
standardization Processes. In Proceedings of MISQ Special Issue Workshop on Standard
Making: A Critical Research Frontier for Information Systems (pp. 92–107).
Dosi, G. (1982). Technological paradigms and technological trajectories - a suggested
interpretation of the determinants and directions of technical change. Research Policy,
11(3), 147–162. http://doi.org/10.1016/0048-7333(82)90016-6
Eisenmann, T., Parker, G., & Alstyne, M. W. Van. (2006). Strategies for two-sided markets.
Harvard Business Review, 84(10). http://doi.org/10.1007/s00199-006-0114-6
Eisenmann, T., Parker, G., & Alstyne, M. W. Van. (2011). Platform envelopment. Strategic
Management Journal, 32, 1270–1285. http://doi.org/10.1002/smj
Erdi, P., Makovi, K., Somogyvari, Z., Strandburg, K., Tobochnik, J., Volf, P., & Zalanyi, L.
(2013). Prediction of emerging technologies based on analysis of the US patent citation
network. Scientometrics, 95(1), 225–242. http://doi.org/10.1007/s11192-012-0796-4
Page 38
37
Erstling, J. (2011). Patent law and the duty of candor: Rethinking the limits of disclosure.
Creighton Law Review, 44, 329–366.
Farber, D., & Baran, P. (1977). The convergence of computing and telecommunications
systems. Science, 195(4283), 1166–70. http://doi.org/10.1126/science.195.4283.1166
Farrell, J. (1989). Standardization and intellectual property. Jurimetrics, 30(1), 35–50.
Fleming, L., & Sorenson, O. (2004). Science as a map in technological search. Strategic
Management Journal, 25(8–9), 909–928. http://doi.org/10.1002/smj.384
Fontana, R., Nuvolari, A., & Verspagen, B. (2009). Mapping technological trajectories as
patent citation networks: An application to data communication standards. Economics of
Innovation and New Technology, 18(4), 311–336. http://doi.org/10.1080/10438590801969073
Fortunato, S. (2010). Community detection in graphs. Physics Reports, 486(3–5), 75–174.
Foster, C., & Heeks, R. (2013). Innovation and scaling of ICT for the bottom-of-the-pyramid.
Journal of Information Technology, 28(4), 296–315. http://doi.org/10.1057/jit.2013.19
Freeman, L. C. (1977). A set of measures of centrality based on betweenness. Sociometry, 40(1),
35–41. http://doi.org/10.2307/3033543
Freeman, L. C. (1979). Centrality in social networks: Conceptual clarification. Social Networks,
1(3), 215–239.
Galloway, A. (2004). Intimations of everyday life: Ubiquitous computing and the city.
Cultural Studies, 18(2), 384–408. http://doi.org/10.1080/0950238042000201572
Gandal, N., Gantman, N., & Genesove, D. (2007). Intellectual property and standardization
committee participation in the US modem industry. In S. Greenstein & V. Stango (Eds.),
Standards and Public Policy (pp. 208–230). Cambridge University Press. Retrieved from
https://books.google.com/books?hl=en&lr=&id=3hMKHwUmaZ8C&oi=fnd&pg=PA208
&dq=gandahl+2007&ots=rx_gAiePHD&sig=rR3ihQL4PCJX2T3xAEnwIwKmh2w
Gauch, S., & Blind, K. (2015). Technological convergence and the absorptive capacity of
standardisation. Technological Forecasting and Social Change, 91, 236–249.
http://doi.org/10.1016/j.techfore.2014.02.022
Gawer, A., & Cusumano, M. A. (2008). How Companies Become Platform Leaders. MIT
Sloan Management Review, 49(2), 28–35. Retrieved from http://sloanreview.mit.edu/the-
magazine/articles/2008/winter/49201/how-companies-become-platform-leaders/
Gay, C., Le Bas, C., Patel, P., & Touach, K. (2005). The determinants of patent citations: an
empirical analysis of French and British patents in the US. Economics of Innovation and
New Technology, 14(5), 339–350. http://doi.org/10.1080/1040859042000307329
Gilsing, V., Nooteboom, B., Vanhaverbeke, W., Duysters, G., & van den Oord, A. (2008).
Network embeddedness and the exploration of novel technologies: Technological
distance, betweenness centrality and density. Research Policy, 37(10), 1717–1731.
http://doi.org/10.1016/j.respol.2008.08.010
Gnyawali, D. R., & Park, B. J. (2011). Co-opetition between giants: Collaboration with
competitors for technological innovation. Research Policy, 40(5), 650–663.
http://doi.org/10.1016/j.respol.2011.01.009
Gomes-Casseres, B. (1994). Group versus group: How alliance networks compete. Harvard
Business Review, 72(4), 62–66.
Greenstein, S., & Khanna, T. (1997). What does industry convergence mean? In D. B. Yoffie
(Ed.), Competing in the Age of Digital Convergence (pp. 201–226). Harvard Business School
Press.
Page 39
38
Grove, A. S. (1996). Only the Paranoid Survive: How to Exploit the Crisis Points that Challenge
Every Company. Currency Doubleday.
Gubbi, J., Buyya, R., Marusic, S., & Palaniswami, M. (2013). Internet of Things (IoT): A vision,
architectural elements, and future directions. Future Generation Computer Systems, 29(7),
1645–1660. http://doi.org/10.1016/j.future.2013.01.010
Hall, B., Jaffe, A., & Trajtenberg, M. (2001). The NBER patent citation data file: Lessons, insights
and methodological tools (No. 8498). National Bureauof Economic Research. Retrieved
from http://www.nber.org/papers/w8498
Han, E. J., & Sohn, S. Y. (2015). Patent valuation based on text mining and survival analysis.
Journal of Technology Transfer, 40(5), 821–839. http://doi.org/10.1007/s10961-014-9367-6
Han, E. J., & Sohn, S. Y. (2016). Technological convergence in standards for information and
communication technologies. Technological Forecasting and Social Change, 106, 1–10.
http://doi.org/10.1016/j.techfore.2016.02.003
Hanseth, O., Monteiro, E., & Hatling, M. (1996). Developing information infrastructure : The
tension between standardization and flexibility. Science, Technology, & Human Values,
21(4), 407–426.
Hargadon, A., & Sutton, R. I. (1997). Technology brokering and innovation in a product
development firm. Administrative Science Quarterly, 42(4), 716.
http://doi.org/10.2307/2393655
Harhoff, D., Scherer, F. M., & Vopel, K. (2003). Citations, family size, opposition and the
value of patent rights. Research Policy, 32(8), 1343–1363. http://doi.org/10.1016/S0048-
7333(02)00124-5
Hausman, J. A., Hall, B. H., & Griliches, Z. (1984). Econometric models for count data with
an application to the patents-R&D relationship. Econometrica, 52(4), 909–938.
Hill, K. (2013, June 20). Seven Networks to work with Intel on mobile network technologies.
RCRWireless. Retrieved from http://www.rcrwireless.com/20130620/telecom-
software/seven-networks-work-intel-mobile-network-technologies
Hughes, T. P. (1983). Networks of Power: Electrification in Western Society, 1880-1930. The Johns
Hopkins University Press.
Hughes, T. P. (1987). The evolution of large technological systems. In W. E. Bijker, T. P.
Hughes, & T. J. Pinch (Eds.), The Social Construction of Technological Systems: New
Directions in the Sociology and History of Technology (pp. 51–82). MIT Press.
Hummon, N. P., & Doreian, P. (1989). Connectivity in a citation network: The development
of DNA theory. Social Networks, 11(1), 39–63.
Jacobides, M. G., Knudsen, T., & Augier, M. (2006). Benefiting from innovation: Value
creation, value appropriation and the role of industry architectures. Research Policy, 35(8
SPEC. ISS.), 1200–1221. http://doi.org/10.1016/j.respol.2006.09.005
Jaffe, A. B., & Trajtenberg, M. (2002). Patents, Citations, and Innovations: A Window on the
Knowledge Economy. The MIT Press.
Kang, B., Huo, D., & Motohashi, K. (2014). Comparison of Chinese and Korean companies in
ICT global standardization: Essential patent analysis. Telecommunications Policy, 38(10),
902–913. http://doi.org/10.1016/j.telpol.2014.09.004
Kang, B., & Motohashi, K. (2015). Essential intellectual property rights and inventors’
involvement in standardization. Research Policy, 44(2), 483–492.
http://doi.org/10.1016/j.respol.2014.10.012
Page 40
39
Karki, M. M. S., & Krishnam, K. S. (1997). Patent citation analysis: A policy analysis tool.
World Patent Information, 19(4), 269–272. http://doi.org/10.1016/S0172-2190(97)00033-1
Kenney, M., & Pon, B. (2011). Structuring the smartphone industry: Is the mobile Internet OS
platform the key? Journal of Industry, Competition and Trade, 11, 239–261.
http://doi.org/10.1007/s10842-011-0105-6
Kim, D., & Lee, H. (2016). Effects of user experience on user resistance to change to the voice
user interface of an in‑vehicle infotainment system: Implications for platform and
standards competition. International Journal of Information Management.
http://doi.org/10.1016/j.ijinfomgt.2016.04.011
Kim, H. (2014). Internet of Things: Concept, Technology, and Business. Hongrung Publishing
Company (in Korean).
Kindleberger, C. P. (1983). Standards as public, collective and private goods. Kyklos, 36(3),
377–396.
Landauer, T. K., Folt, P. W., & Laham, D. (1998). An Introduction to Latent Semantic
Analysis. Discourse Processes, 25, 259–284. http://doi.org/10.1080/01638539809545028
Lee, K., & Lim, C. (2001). Technological regimes, catching-up and leapfrogging: findings
from the Korean industries. Research Policy, 30(3), 459–483. http://doi.org/10.1016/S0048-
7333(00)00088-3
Lee, K., Lim, C., & Song, W. (2005). Emerging digital technology as a window of opportunity
and technological leapfrogging: catch-up in digital TV by the Korean firms. International
Journal of Technology Management, 29(1/2), 40. http://doi.org/10.1504/IJTM.2005.006004
Lee, M. D., Pincombe, B., & Welsh, M. (2005). An empirical evaluation of models of text
document similarity. In B. G. Bara, L. Barsalou, & M. Bucciarelli (Eds.), XXVII Annual
Conference of the Cognitive Science Society (pp. 1254–1259). Cognitive Science Society.
Retrieved from http://digital.library.adelaide.edu.au/dspace/handle/2440/28910
Leiponen, A. (2008). Competing through cooperation: The organization of standard setting
in wireless telecommunications. Management Science, 54(11), 1904–1919.
http://doi.org/10.1287/mnsc.1080.0912
Lemko. (2013). Lemko Corporation Launches 100% Virtualized, No Core, M2M Quick-
Deploy-PlatformTM. Retrieved December 6, 2016, from
http://www.prnewswire.com/news-releases/lemko-corporation-launches-100-
virtualized-no-core-m2m-quick-deploy-platform-207691201.html
Lemley, M. (2002). Intellectual property rights and standard-setting organizations. California
Law Review, 90, 1889–1980. Retrieved from http://www.jstor.org/stable/3481437
Leydesdorff, L., de Moya-Anegón, F., & Guerrero-Bote, V. P. (2015). Journal maps,
interactive overlays, and the measurement of interdisciplinarity on the basis of scopus
data (1996–2012). Journal of the Associtation for Information Science and Technology, 66(5),
1001–1016. http://doi.org/10.1002/asi
Martinelli, A. (2012). An emerging paradigm or just another trajectory? Understanding the
nature of technological changes using engineering heuristics in the telecommunications
switching industry. Research Policy, 41(2), 414–429.
http://doi.org/10.1016/j.respol.2011.10.012
Mason, C. H., & Perreault, W. D. (1991). Collinearity, power, and interpretation of multiple
regression analysis. Journal of Marketing Research, 38, 268–280.
Page 41
40
McGahan, A., Vadasz, L. L., & Yoffie, D. B. (1997). Creating value and setting standards: The
lessons of consumer electronics for personal digital assistants. In D. B. Yoffie (Ed.),
Competing in the Age of Digital Convergence (pp. 227–264). Harvard Business School Press.
Metcalfe, J. S., & Miles, I. (1994). Standards, selection and variety: an evolutionary approach.
Information Economics and Policy, 6(3–4), 243–268. http://doi.org/10.1016/0167-
6245(94)90004-3
Meyer, M. (2000). Does science push technology? Patents citing scientific literature. Research
Policy, 29, 409–434.
Mokyr, J. (2002). The Gifts of Athena: Historical Origins of the Knowledge Economy. Princeton
Unversity Press.
Mu, Q., & Lee, K. (2005). Knowledge diffusion, market segmentation and technological
catch-up: The case of the telecommunication industry in China. Research Policy, 34(6),
759–783. http://doi.org/10.1016/j.respol.2005.02.007
Narin, F., Hamilton, K. S., & Olivastro, D. (1997). The incresing linkage between U.S.
technology and public science. Research Policy, 26, 317–330.
Nelson, R. R., & Winter, S. G. (1977). In search of a useful theory of innovation. Research
Policy, 6, 36–76.
Nelson, R. R., & Winter, S. G. (1982). An Evolutionary Theory of Economic Change. Harvard
University Press.
Newman, M. E. J. (2006). Modularity and community structure in networks. In Proceedings of
the National Academy of Sciences (Vol. 103, pp. 8577–8582).
Nohria, N., & Garcia-Pont, C. (1991). Global strategic linkages and industry structure.
Strategic Management Journal, 12(Special Issue: Global Strategy), 105–124.
http://doi.org/10.1002/smj.4250120909
Norp, T., & Landais, B. (2012). M2M optimizations and public mobile networks. In D.
Boswarthick, O. Elloumi, & O. Hersent (Eds.), M2M Communications: A Systems
Approach (pp. 148–192). John Wiley & Sons.
Park, K. H., & Lee, K. (2006). Linking the technological regime to the technological catch-up:
Analyzing Korea and Taiwan using the US patent data. Industrial and Corporate Change,
15(4), 715–753. http://doi.org/10.1093/icc/dtl016
Pasquinelli, M. (2009). Google’s PageRank algorithm: A diagram of cognitive capitalism and
the rentier of the common intellect. In K. Becker & F. Stalder (Eds.), Deep Search: The
Politics of Search Beyond Google (pp. 152–162). London: Transaction Publishers.
Patel, P., & Pavitt, K. (1997). The technological competencies of the world’s largest firms:
Complex and path-dependent, but not much variety. Research Policy, 26(2), 141–156.
http://doi.org/10.1016/S0048-7333(97)00005-X
Perez, C., & Soete, L. (1988). Catching up in technology: entry barriers and windows of
opportunity. In C. Freeman, R. Nelson, G. Silverberg, & L. Soete (Eds.), Technical Change
and Economic Theory (pp. 458–479). London: Pinter.
Pinch, T. J., & Bijker, W. E. (1987). The social construction of facts and artifacts: Or how the
sociology of science and the sociology of technology might benefit each other. In W. E.
Bijker, T. P. Hughes, & T. J. Pinch (Eds.), The Social Construction of Technological Systems:
New Directions in the Sociology and History of Technology (pp. 17–50). MIT Press.
Podolny, J. M., & Stuart, T. E. (1995). A role-based ecology of technological change. American
Journal of Sociology, 100(5), 1224–1260. http://doi.org/10.1086/230637
Page 42
41
Porter, M. E., & Heppelmann, J. E. (2014). How smart, connected products are transforming
competition. Harvard Business Review, (November), 65–88.
Powell, J. L. (1984). Least absolute deviations estimation for the censored regression model.
Journal of Econometrics, 25(3), 303–325. http://doi.org/10.1016/0304-4076(84)90004-6
Rice, J., & Galvin, P. (2006). Alliance patterns during industry life cycle emergence: The case
of Ericsson and Nokia. Technovation, 26(3), 384–395.
http://doi.org/10.1016/j.technovation.2005.02.005
Rosenberg, N. (1969). The direction of technological change: Inducement mechanisms and
focusing devices. Economic Development and Cultural Change, 18(1), 1–24.
Rosenberg, N. (1976). Perspectives on Technology. Cambridge University Press.
Rysman, M., & Simcoe, T. (2008). Patents and the performance of voluntary standard-setting
organizations. Management Science, 54(11), 1920–1934.
http://doi.org/10.1287/mnsc.1080.0919
Sadowski, B. M., Dittrich, K., & Duysters, G. M. (2003). Collaborative strategies in the event
of technological discontinuities: The case of Nokia in the mobile telecommunication
industry. Small Business Economics, 21(2), 173–186.
Sampat, B. N., & Ziedonis, A. (2004). Patent citations and the economic value of patents. In
H. F. Moed, W. Glanzel, & U. Schmoch. (Eds.), The Handbook of Quantitative Science and
Technology Research (pp. 277–298). Kluwer Academic Publishers.
Samuelson, P. (1954). The pure theory of public expenditure. The Review of Economics and
Statistics. Retrieved from http://www.jstor.org/stable/1925895
Saviotti, P. P., & Metcalfe, J. S. (1984). A theoretical approach to the construction of
technological output indicators. Research Policy, 13(3), 141–151.
http://doi.org/10.1016/0048-7333(84)90022-2
Schmidt, S. K., & Werle, R. (1998). Coordinating technology: Studies in the international
standardization of telecommunications. MIT Press.
Schneiderman, R. (2015). Modern Standardization: Case Studies at the Crossroads of Technology,
Economics, and Politics. John Wiley & Sons.
Schoenberger, C. R. (2002, March). The internet of things. Forbes Magazine. Retrieved from
http://www.forbes.com/global/2002/0318/092.html
Schumpeter, J. (1950). Capitalism, Socialism, and Democracy (3rd ed.). Harper Perennial
Modern Classics (originally published in 1942).
Shapiro, C., & Varian, H. R. (1999a). Information Rules: A Strategic Guide to the Network
Economy. Harvard Business School Press.
Shapiro, C., & Varian, H. R. (1999b). The art of standards wars. California Management Review,
41(2), 8–32.
Sharma, S., Durand, R., & Gur-Arie, O. (1981). Identification and analysis of moderator
variables. Journal of Marketing Research, 18, 291–300. Retrieved from
http://www.jstor.org/stable/3150970
Shen, S., & Carug, M. (2014). An evolutionary way to standardize the Internet of Things.
Journal of ICT Standardization, 2(2), 87–108. http://doi.org/10.13052/jicts2245-800X.222
Shin, D. (2014). A socio-technical framework for Internet-of-Things design: A human-
centered design for the Internet of Things. Telematics and Informatics, 31(4), 519–531.
http://doi.org/10.1016/j.tele.2014.02.003
Page 43
42
Stuart, T. E. (1998). Network positions and propensities to collaborate: An investigation of
strategic alliance formation in a high-technology industry. Administrative Science
Quarterly, 43(3), 668–698. http://doi.org/10.2307/2393679
Suarez, F. F. (2004). Battles for technological dominance: An integrative framework. Research
Policy, 33(July 2003), 271–286. http://doi.org/10.1016/j.respol.2003.07.001
The Economist. (2014). Google and the internet of things: Feathering its Nest. The Economist.
Tijssen, R. J. W. (2001). Global and domestic utilization of industrial relevant science: Patent
citation analysis of science-technology interactions and knowledge flows. Research
Policy, 30(1), 35–54. http://doi.org/10.1016/S0048-7333(99)00080-3
Trajtenberg, M., Henderson, R., & Jaffe, A. B. (2002). University versus corporate patents: A
window on the basicness of invention. In A. B. Jaffe & M. Trajtenberg (Eds.), Patents,
Citations, and Innovations: A Window on the Knowledge Economy (pp. 51–87). The MIT
Press.
Verspagen, B. (2007). Mapping technological trajectories as patent citation networks: A
study on the history of fuel cell research. Advances in Complex Systems, 10(1), 93–115.
Ward, J. H. (1963). Hierarchical grouping to optimize an objective function. Journal of the
American Statistical Association, 58(301), 236–244.
http://doi.org/10.1080/01621459.1963.10500845
Wasserman, S., & Faust, K. (1994). Social Network Analysis: Methods and Applications.
Cambridge University Press.
Yoffie, D. B. (Ed.). (1997a). Competing in the Age of Digital Convergence. Boston: Harvard
Business School Press.
Yoffie, D. B. (1997b). Introduction: CHESS and competing in the age of digital convergence.
In D. B. Yoffie (Ed.), Competing in the Age of Digital Convergence (pp. 1–35). Harvard
Business School Press.
Yu, J. (2011). From 3G to 4G: technology evolution and path dynamics in China’s mobile
telecommunication sector. Technology Analysis & Strategic Management, 23(10), 1079–
1093. http://doi.org/10.1080/09537325.2011.621306
Zhou, H. (2012). The Internet of Things in the Cloud: A Middleware Perspective. CRC Press.
Zoo, H., Vries, H. J. De, & Lee, H. (2017). Interplay of innovation and standardization:
Exploring the relevance in developing countries. Technological Forecasting & Social
Change, 118, 334–348. http://doi.org/10.1016/j.techfore.2017.02.033
Appendix
1. Clusters in the M2M/IoT patent citation network, identified by the Chinese
Whispers algorithm
Page 44
43
Clusters 1–10 similarly matches communities 1–10, detected by Blondel et al.’s modularity
optimization method, as shown in Figure 2. The size of nodes is measured based on
betweenness centrality.
2. Glossary
3GPP 3rd generation partnership project
CLAD Censored least absolute deviations
CPC Cooperative patent classification
ETSI European telecommunications standards institute
GSM Global system for mobile communication
IEEE Institute of electrical and electronics engineers
IETF Internet engineering task force
IIC Industrial internet consortium
IPC International patent classification
IoT Internet of things
LSA Latent semantic analysis
M2M Machine-to-machine
MTC Machine-type communications
NPL Non-patent literature
OIC Open interconnect consortium
RFID Radio-frequency identification
SCADA Supervisory control and data acquisition
SSO Standards-setting organization
Page 45
44
STD Standards
TR Technical report
TS Technical specification
UMTS Universal mobile telecommunications service
VIF Variance influence factors
WCDMA Wideband code division multiple access
WSN Wireless sensor network
WiMax Worldwide interoperability for microwave access
